MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM

0Associated
Cases 
0Associated
Defendants 
0Accused
Products 
143Forward
Citations 
0
Petitions 
2
Assignments
First Claim
1. A method of providing a modulated signal useable in a signal transmission system, the method comprising:
 transforming a data frame including a plurality of data elements into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements; and
generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.
2 Assignments
0 Petitions
Accused Products
Abstract
A system and method of providing a modulated signal useable in a signal transmission system. The method includes transforming, perhaps with respect to both time and frequency, a data frame including a plurality of data elements into a transformed data matrix. The transformed data matrix includes a plurality of transformed data elements where each of the plurality of transformed data elements is based upon each of the plurality of data elements. The method further includes generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.
153 Citations
MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM  
Patent #
US 20140169437A1
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,031,141 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,071,286 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,071,285 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,130,638 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,294,315 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,548,840 B2
Filed 05/20/2015

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,660,851 B2
Filed 06/29/2015

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,712,354 B2
Filed 07/24/2015

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,729,281 B2
Filed 05/11/2015

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

System and method for coordinated management of network access points  
Patent #
US 9,866,363 B2
Filed 06/20/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

System and method for implementing orthogonal time frequency space communications using OFDM  
Patent #
US 9,893,922 B2
Filed 02/17/2017

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 9,900,048 B2
Filed 09/14/2015

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Orthogonal time frequency space communication system compatible with OFDM  
Patent #
US 9,912,507 B2
Filed 06/27/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Orthogonal time frequency space modulation system  
Patent #
US 9,929,783 B2
Filed 05/11/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Multiple access in an orthogonal time frequency space communication system  
Patent #
US 9,967,758 B2
Filed 09/08/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Shlomo Selim Rakib, Ronny Hadani

Symplectic orthogonal time frequency space modulation system  
Patent #
US 10,003,487 B2
Filed 06/21/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modular antennas with heat sinks in wireless power transmission systems  
Patent #
US 10,008,886 B2
Filed 03/03/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power transmitter configured to transmit power waves to a predicted location of a moving wireless power receiver  
Patent #
US 10,008,875 B1
Filed 09/16/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Method for automatically testing the operational status of a wireless power receiver in a wireless power transmission system  
Patent #
US 10,008,889 B2
Filed 08/21/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power receiver having a charger system for enhanced power delivery  
Patent #
US 10,014,728 B1
Filed 06/23/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for selecting antennas to generate and transmit power transmission waves  
Patent #
US 10,020,678 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Signal separation in an orthogonal time frequency space communication system using MIMO antenna arrays  
Patent #
US 10,020,854 B2
Filed 12/09/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Shlomo Selim Rakib, Ronny Hadani

Proximity transmitters for wireless power charging systems  
Patent #
US 10,021,523 B2
Filed 12/24/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Antenna for transmitting wireless power signals  
Patent #
US 10,027,159 B2
Filed 03/03/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

3D triple linear antenna that acts as heat sink  
Patent #
US 10,027,180 B1
Filed 11/02/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Near field transmitters for wireless power charging of an electronic device by leaking RF energy through an aperture  
Patent #
US 10,027,158 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for generating and transmitting wireless power transmission waves using antennas having a spacing that is selected by the transmitter  
Patent #
US 10,027,168 B2
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for determining and generating a waveform for wireless power transmission waves  
Patent #
US 10,033,222 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods of wireless power charging through multiple receiving devices  
Patent #
US 10,038,332 B1
Filed 02/19/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power supply for rescue devices  
Patent #
US 10,038,337 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power transmission device having antennas oriented in three dimensions  
Patent #
US 10,050,470 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Methods and systems for maximum power point transfer in receivers  
Patent #
US 10,056,782 B1
Filed 04/10/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Stamped threedimensional antenna  
Patent #
US 10,063,108 B1
Filed 11/02/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and method for a selfsystem analysis in a wireless power transmission network  
Patent #
US 10,063,106 B2
Filed 05/23/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Proximity transmitters for wireless power charging systems  
Patent #
US 10,063,105 B2
Filed 12/24/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

TomlinsonHarashima precoding in an OTFS communication system  
Patent #
US 10,063,295 B2
Filed 04/03/2017

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 10,063,354 B2
Filed 01/17/2017

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

System and method for generating a power receiver identifier in a wireless power network  
Patent #
US 10,063,064 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Integrated miniature PIFA with artificial magnetic conductor metamaterials  
Patent #
US 10,068,703 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

External or internal wireless power receiver with spacedapart antenna elements for charging or powering mobile devices using wirelessly delivered power  
Patent #
US 10,075,017 B2
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Nearfield RF charging pad with multiband antenna element with adaptive loading to efficiently charge an electronic device at any position on the pad  
Patent #
US 10,079,515 B2
Filed 12/06/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Multiple access in an orthogonal time frequency space communication system  
Patent #
US 10,090,973 B2
Filed 09/19/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Wireless powered house  
Patent #
US 10,090,699 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and method for enabling automatic charging schedules in a wireless power network to one or more devices  
Patent #
US 10,090,886 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and method for twodimensional equalization in an orthogonal time frequency space communication system  
Patent #
US 10,090,972 B2
Filed 09/07/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Shlomo Selim Rakib, Ronny Hadani

Protocols for authenticated wireless power transmission  
Patent #
US 10,103,552 B1
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Transmitters for wireless power transmission  
Patent #
US 10,103,582 B2
Filed 05/10/2013

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Near field transmitters with harmonic filters for wireless power charging  
Patent #
US 10,116,162 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Methods and systems for maximum power point transfer in receivers  
Patent #
US 10,116,170 B1
Filed 09/02/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Integrated antenna arrays for wireless power transmission  
Patent #
US 10,116,143 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for assigning a set of antennas of a wireless power transmitter to a wireless power receiver based on a location of the wireless power receiver  
Patent #
US 10,122,415 B2
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems, methods, and devices for using a battery as a antenna for receiving wirelessly delivered power from radio frequency power waves  
Patent #
US 10,122,219 B1
Filed 04/12/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless charging and powering of electronic sensors in a vehicle  
Patent #
US 10,124,754 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for identifying receiver locations using sensor technologies  
Patent #
US 10,128,686 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and method for providing health safety in a wireless power transmission system  
Patent #
US 10,128,693 B2
Filed 07/14/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods of providing wireless power using receiver device sensor inputs  
Patent #
US 10,128,699 B2
Filed 03/03/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for preconfiguring transmission devices for power wave transmissions based on location data of one or more receivers  
Patent #
US 10,135,294 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Near field transmitters for wireless power charging of an electronic device by leaking RF energy through an aperture offset from a patch antenna  
Patent #
US 10,135,286 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

3D antenna mount  
Patent #
US 10,135,112 B1
Filed 11/02/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for nullifying energy levels for wireless power transmission waves  
Patent #
US 10,135,295 B2
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for maximizing wireless power transfer efficiency by instructing a user to change a receiver device's position  
Patent #
US 10,141,768 B2
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Near field transmitters with contact points for wireless power charging  
Patent #
US 10,141,771 B1
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for controlling communications during wireless transmission of power using application programming interfaces  
Patent #
US 10,141,791 B2
Filed 05/07/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power transmission with selective range  
Patent #
US 10,148,133 B2
Filed 07/20/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for using a predetermined number of communication channels of a wireless power transmitter to communicate with different wireless power receivers  
Patent #
US 10,148,097 B1
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for designating a master power transmitter in a cluster of wireless power transmitters  
Patent #
US 10,153,645 B1
Filed 03/04/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for using application programming interfaces to control communications between a transmitter and a receiver  
Patent #
US 10,153,653 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for preconfiguring sensor data for wireless charging systems  
Patent #
US 10,153,660 B1
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for identifying receivers in a transmission field by transmitting exploratory power waves towards different segments of a transmission field  
Patent #
US 10,158,259 B1
Filed 09/16/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for symplectic orthogonal time frequency shifting modulation and transmission of data  
Patent #
US 10,158,394 B2
Filed 05/11/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

System and methods for using sound waves to wirelessly deliver power to electronic devices  
Patent #
US 10,158,257 B2
Filed 01/03/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Modular antenna boards in wireless power transmission systems  
Patent #
US 10,164,478 B2
Filed 03/03/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for managing and controlling a wireless power network by establishing time intervals during which receivers communicate with a transmitter  
Patent #
US 10,170,917 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Radiating metamaterial antenna for wireless charging  
Patent #
US 10,177,594 B2
Filed 12/26/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Receiver device with antennas positioned in gaps  
Patent #
US 10,186,892 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for real time or near real time wireless communications between a wireless power transmitter and a wireless power receiver  
Patent #
US 10,186,893 B2
Filed 03/04/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and methods for pocketforming based on constructive and destructive interferences to power one or more wireless power receivers using a wireless power transmitter including a plurality of antennas  
Patent #
US 10,186,913 B2
Filed 12/27/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Cluster management of transmitters in a wireless power transmission system  
Patent #
US 10,193,396 B1
Filed 03/04/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for wirelessly transmitting power from a transmitter to a receiver by determining refined locations of the receiver in a segmented transmission field associated with the transmitter  
Patent #
US 10,199,850 B2
Filed 09/16/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Radar motion detection using stepped frequency in wireless power transmission system  
Patent #
US 10,199,835 B2
Filed 03/02/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Method for automatically testing the operational status of a wireless power receiver in a wireless power transmission system  
Patent #
US 10,199,849 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Compact PIFA antenna  
Patent #
US 10,205,239 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and methods for wireless power transmission to an electronic device in accordance with userdefined restrictions  
Patent #
US 10,206,185 B2
Filed 06/03/2013

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for real or near real time wireless communications between a wireless power transmitter and a wireless power receiver  
Patent #
US 10,211,685 B2
Filed 03/04/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Method for 3 dimensional pocketforming  
Patent #
US 10,211,680 B2
Filed 07/19/2013

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for controlling operation of a transmitter of a wireless power network based on user instructions received from an authenticated computing device powered or charged by a receiver of the wireless power network  
Patent #
US 10,211,682 B2
Filed 05/07/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless charging using selected reflectors  
Patent #
US 10,211,674 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Receiver chip for routing a wireless signal for wireless power charging or data reception  
Patent #
US 10,218,207 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Compact PIFA antenna  
Patent #
US 10,218,227 B2
Filed 05/07/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for paymentbased authorization of wireless power transmission service  
Patent #
US 10,223,717 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless powering of electronic devices with selective delivery range  
Patent #
US 10,224,758 B2
Filed 11/01/2013

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power receivers that communicate status data indicating wireless power transmission effectiveness with a transmitter using a builtin communications component of a mobile device, and methods of use thereof  
Patent #
US 10,230,266 B1
Filed 12/31/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wearable device with wireless power and payload receiver  
Patent #
US 10,243,414 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Antenna having coaxial structure for near field wireless power charging  
Patent #
US 10,256,657 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Nearfield RF charging pad with adaptive loading to efficiently charge an electronic device at any position on the pad  
Patent #
US 10,256,677 B2
Filed 02/03/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Transmitter board allowing for modular antenna configurations in wireless power transmission systems  
Patent #
US 10,263,476 B2
Filed 03/03/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Multimode transmitter with an antenna array for delivering wireless power and providing WiFi access  
Patent #
US 10,263,432 B1
Filed 12/30/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods of object detection in wireless power charging systems  
Patent #
US 10,270,261 B2
Filed 02/25/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Nearfield charging pad for wireless power charging of a receiver device that is temporarily unable to communicate  
Patent #
US 10,277,054 B2
Filed 02/17/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods of controlling transmission of wireless power based on object indentification using a video camera  
Patent #
US 10,291,056 B2
Filed 02/25/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Power transmission control systems and methods  
Patent #
US 10,291,066 B1
Filed 12/31/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for controlling farfield wireless power transmission based on battery power levels of a receiving device  
Patent #
US 10,291,055 B1
Filed 12/29/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power transmitter that selectively activates antenna elements for performing wireless power transmission  
Patent #
US 10,291,294 B2
Filed 11/21/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless power transmitters for selecting antenna sets for transmitting wireless power based on a receiver's location, and methods of use thereof  
Patent #
US 10,298,024 B2
Filed 12/12/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Synchronous rectifier design for wireless power receiver  
Patent #
US 10,298,133 B2
Filed 01/02/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for wireless charging using a cordless transceiver  
Patent #
US 10,305,315 B2
Filed 11/07/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for wireless power charging  
Patent #
US 10,312,715 B2
Filed 09/16/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Miniaturized highlyefficient designs for nearfield power transfer system  
Patent #
US 10,320,446 B2
Filed 09/19/2016

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Crosspolarized dipole antenna  
Patent #
US 10,333,332 B1
Filed 10/13/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

OTFS methods of data channel characterization and uses thereof  
Patent #
US 10,334,457 B2
Filed 01/04/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Shlomo Selim Rakib, Ronny Hadani

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 10,341,155 B2
Filed 05/22/2017

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Integrated circuit for managing wireless power transmitting devices  
Patent #
US 10,355,534 B2
Filed 04/26/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Iterative two dimensional equalization of orthogonal time frequency space modulated signals  
Patent #
US 10,355,887 B2
Filed 10/01/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Variable beamwidth multiband antenna  
Patent #
US 10,356,632 B2
Filed 01/27/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Integrated antenna structure arrays for wireless power transmission  
Patent #
US 10,381,880 B2
Filed 07/21/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Surface mount dielectric antennas for wireless power transmitters  
Patent #
US 10,389,161 B2
Filed 03/15/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Receiver for wireless power reception having a backup battery  
Patent #
US 10,396,588 B2
Filed 01/16/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for operating a plurality of antennas of a wireless power transmitter  
Patent #
US 10,396,604 B2
Filed 12/26/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Orthogonal time frequency space communication system compatible with OFDM  
Patent #
US 10,411,843 B2
Filed 07/12/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Systems and methods for automatically testing the communication between wireless power transmitter and wireless power receiver  
Patent #
US 10,439,448 B2
Filed 08/21/2014

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Microstrip antennas for wireless power transmitters  
Patent #
US 10,439,442 B2
Filed 01/24/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Nearfield antenna for wireless power transmission with four coplanar antenna elements that each follows a respective meandering pattern  
Patent #
US 10,447,093 B2
Filed 06/29/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Orthogonal time frequency space modulation system for the Internet of Things  
Patent #
US 10,469,215 B2
Filed 04/29/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Methods of selectively activating antenna zones of a nearfield charging pad to maximize wireless power delivered to a receiver  
Patent #
US 10,476,312 B2
Filed 04/02/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Variable latency data communication using orthogonal time frequency space modulation  
Patent #
US 10,476,564 B2
Filed 12/09/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Systems and methods of object detection using one or more sensors in wireless power charging systems  
Patent #
US 10,483,768 B2
Filed 01/16/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Antenna structures having planar inverted Fantenna that surrounds an artificial magnetic conductor cell  
Patent #
US 10,490,346 B2
Filed 09/04/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Antenna with electromagnetic band gap ground plane and dipole antennas for wireless power transfer  
Patent #
US 10,491,029 B2
Filed 07/17/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Systems and methods for wirelessly delivering power to electronic devices in response to commands received at a wireless power transmitter  
Patent #
US 10,498,144 B2
Filed 10/10/2017

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Slot antenna with orthogonally positioned slot segments for receiving electromagnetic waves having different polarizations  
Patent #
US 10,511,196 B2
Filed 07/17/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Nearfield antennas for accumulating energy at a nearfield distance with minimal farfield gain  
Patent #
US 10,511,097 B2
Filed 05/08/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Unit cell of a wireless power transmitter for wireless power charging  
Patent #
US 10,516,289 B2
Filed 07/31/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

System and methods for using sound waves to wirelessly deliver power to electronic devices  
Patent #
US 10,516,301 B2
Filed 12/18/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless charging transmitters that use sensor data to adjust transmission of power waves  
Patent #
US 10,523,058 B2
Filed 08/28/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Receiver devices configured to determine location within a transmission field  
Patent #
US 10,523,033 B2
Filed 09/22/2015

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

TomlinsonHarashima precoding in an OTFS communication system  
Patent #
US 10,541,734 B2
Filed 07/24/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Systems and methods for determining when to transmit power waves to a wireless power receiver  
Patent #
US 10,554,052 B2
Filed 01/30/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

Wireless telecommunications system for highmobility applications  
Patent #
US 10,555,281 B2
Filed 03/31/2017

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Windowed sequence for random access method and apparatus  
Patent #
US 10,568,143 B2
Filed 03/27/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 10,567,125 B2
Filed 07/24/2018

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

System and method for providing wireless communication services using configurable broadband infrastructure shared among multiple network operators  
Patent #
US 10,574,317 B2
Filed 06/20/2016

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Stamped threedimensional antenna  
Patent #
US 10,594,165 B2
Filed 08/28/2018

Current Assignee
Energous Corporation

Sponsoring Entity
Energous Corporation

FORWARD SUBSTITUTION FOR ERRORCORRECTION ENCODING AND THE LIKE  
Patent #
US 20110131463A1
Filed 12/22/2009

Current Assignee
Avago Technologies International Sales Pte Limited

Sponsoring Entity
Avago Technologies International Sales Pte Limited

Configurable system for performing repetitive actions and method for configuring and operating same  
Patent #
US 20070078661A1
Filed 11/14/2006

Current Assignee
PortalPlayer USA Inc.

Sponsoring Entity
PortalPlayer USA Inc.

Apparatus and method for the compression and transmission of multiformat data  
Patent #
US 5,182,642 A
Filed 04/19/1991

Current Assignee
General Dynamics Land Systems Incorporated

Sponsoring Entity
General Dynamics Land Systems Incorporated

MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM  
Patent #
US 20140161154A1
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM  
Patent #
US 20140169436A1
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM  
Patent #
US 20140169437A1
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM  
Patent #
US 20140169406A1
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Orthonormal timefrequency shifting and spectral shaping communications method  
Patent #
US 8,879,378 B2
Filed 05/26/2011

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Modulation and equalization in an orthonormal timefrequency shifting communications system  
Patent #
US 8,976,851 B2
Filed 06/25/2013

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

Orthonormal timefrequency shifting and spectral shaping communications method  
Patent #
US 9,083,483 B1
Filed 07/27/2014

Current Assignee
Cohere Technologies Inc

Sponsoring Entity
Cohere Technologies Inc

34 Claims
 1. A method of providing a modulated signal useable in a signal transmission system, the method comprising:
transforming a data frame including a plurality of data elements into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements; and generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.  View Dependent Claims (2, 3, 4, 5)
 6. A method of providing a modulated signal useable in a signal transmission system, the method comprising:
transforming a data frame including a plurality of data elements into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements; and generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.  View Dependent Claims (7, 8, 9, 10, 11)
 12. A signal transmitter for use in a communication system, the signal transmitter comprising:
an input port; an output port; a processor; a memory including program code executable by the processor, the program code including; code for receiving, at the input port, a data frame including a plurality of data elements; code for transforming the data frame into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements; and code for generating a modulated signal in accordance with the transformed data elements of the transformed data matrix and for providing the modulated signal to the output port.  View Dependent Claims (13, 14, 15, 16, 34)
 17. A signal transmitter for use in a communication system, the signal transmitter comprising:
an input port; an output port; a processor; a memory including program code executable by the processor, the program code including; code for receiving, at the input port, a data frame including a plurality of data elements; code for transforming the data frame into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements; code for generating a modulated signal in accordance with the transformed data elements of the transformed data matrix.  View Dependent Claims (18, 19, 20, 21, 22, 24)
 23. A nontransitory computer readable medium including program instructions for execution by a processor in a signal transmitter, the program instructions comprising instructions for causing the processor to:
receive, at an input port of the signal transmitter, a data frame including a plurality of data elements; transform the data frame into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements; and generate a modulated signal in accordance with the transformed data elements of the transformed data matrix.  View Dependent Claims (25, 26, 27)
 28. A nontransitory computer readable medium including program instructions for execution by a processor in a signal transmitter, the program instructions comprising instructions for causing the processor to:
receive, at an input port of the signal transmitter, a data frame including a plurality of data elements; transform the data frame into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements; and generate a modulated signal in accordance with the transformed data elements of the transformed data matrix.  View Dependent Claims (29, 30, 31, 32, 33)
1 Specification
The present application claims the benefit of priority under 35 U.S.C. §119(e) of U.S. Provisional Application Ser. No. 61/664,020, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Jun. 25, 2012, of U.S. Provisional Application Ser. No. 61/801,398, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, of U.S. Provisional Application Ser. No. 61/801,366, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, of U.S. Provisional Application Ser. No. 61/801,435, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, of U.S. Provisional Application Ser. No. 61/801,495, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, of U.S. Provisional Application Ser. No. 61/801,994, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, and of U.S. Provisional Application Ser. No. 61/801,968, entitled MODULATION AND EQUALIZATION IN AN ORTHONORMAL TIMEFREQUENCY SHIFTING COMMUNICATIONS SYSTEM, filed Mar. 15, 2013, the contents of each of which are hereby incorporated by reference in their entirety for all purposes. This application is a continuationinpart of U.S. patent application Ser. No. 13/117,119, entitled ORTHONORMAL TIMEFREQUENCY SHIFTING AND SPECTRAL SHAPING COMMUNICATIONS METHOD, filed May 26, 2011, which claims priority to U.S. Provisional Application Ser. No. 61/349,619, entitled ORTHONORMAL TIMEFREQUENCY SHIFTING AND SPECTRAL SHAPING COMMUNICATIONS METHOD, filed May 28, 2010, and is a continuationinpart of U.S. patent application Ser. No. 13/117,124, entitled COMMUNICATIONS METHOD EMPLOYING ORTHONORMAL TIMEFREQUENCY SHIFTING AND SPECTRAL SHAPING, filed May 26, 2011, which claims priority to U.S. Provisional Application Ser. No. 61/349,619, entitled ORTHONORMAL TIMEFREQUENCY SHIFTING AND SPECTRAL SHAPING COMMUNICATIONS METHOD, filed May 28, 2010, the contents of each of which are hereby incorporated by reference in their entirety for all purposes.
This disclosure generally relates communications protocols and methods, and more particularly relates to methods for modulation and related processing of signals used for wireless and other forms of communication.
Modern electronic communication devices, such as devices configured to communicate over transmission media such as optical fibers, electronic wires/cables, or wireless links, all operate by modulating signals and sending these signals over the applicable transmission medium. These signals, which generally travel at or near the speed of light, can be subjected to various types of degradation or channel impairments. For example, echo signals can potentially be generated by optical fiber or wire/cable mediums whenever the modulated signal encounters junctions in the optical fiber or wire/cable. Echo signals can also potentially be generated when wireless signals bounce off of wireless reflecting surfaces, such as the sides of buildings, and other structures. Similarly, frequency shifts can occur when the optical fiber or wire/cable pass through different regions of fiber or cable with somewhat different signal propagating properties or different ambient temperatures. For wireless signals, signals transmitted to or from a moving vehicle can encounter Doppler effects that also result in frequency shifts. Additionally, the underlying equipment (i.e. transmitters and receivers) themselves do not always operate perfectly, and can produce frequency shifts as well.
These echo effects and frequency shifts are undesirable and, if such shifts become too large, may adversely affect network performance by effectively lowering maximum attainable data rates and/or increasing error rates. Such performance degradation is particularly problematic in wireless networks, which are straining to accommodate more and more users, each desiring to send and receive everincreasing amounts of data. Within wireless networks, the adverse effects arising from echo effects and frequency shifts stem at least in part from the characteristics of existing wireless devices having wireless communication capability. In particular, these portable wireless devices (such as cell phones, portable computers, and the like) are often powered by small batteries, and the users of such devices typically expect to operate them for many hours before recharging is required. To meet these user expectations, the wireless transmitters on these devices must output wireless signals using very small amounts of power, making it difficult to distinguish the wireless radio signal over background noise.
An additional problem is that many of these devices are carried on moving vehicles, such as automobiles. This causes additional complications because the lowpower wireless signal transmitted by these devices can also be subjected to various distortions, such as varying and unpredictable Doppler shifts, and unpredictable multipath effects often caused by varying radio reflections off of buildings or other structures.
Moreover, the noise background of the various wireless channels becomes ever higher as noiseproducing electrical devices proliferate. The proliferation of other wireless devices also adds to the background noise.
The system and method for wideband communications disclosed herein is capable of operating using relatively low amounts of power while maintaining improved resistance to problems of Doppler shift, multipath reflections, and background noise. Although examples in the context of wireless communications will be used throughout this application, the methods disclosed herein are intended, unless stated otherwise, to be equally applicable to wired communication systems.
In one aspect, the present disclosure describes a method of providing a modulated signal useable in a signal transmission system. The method of this aspect includes transforming a data frame including a plurality of data elements into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements. This method further includes generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.
In another aspect, the present disclosure describes a method of providing a modulated signal useable in a signal transmission system. The method of this aspect includes transforming a data frame including a plurality of data elements into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements, and generating the modulated signal in accordance with the transformed data elements of the transformed data matrix.
Another aspect of the present disclosure relates to a signal transmitter for use in a communication system. The signal transmitter includes an input port, an output port, a processor and a memory including program code executable by the processor. The program code includes code for receiving, at the input port, a data frame including a plurality of data elements. The program code further includes code for transforming the data frame into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements. In addition, the program code includes code for generating a modulated signal in accordance with the transformed data elements of the transformed data matrix and for providing the modulated signal to the output port.
In yet a further aspect, the present disclosure pertains to a signal transmitter which includes an input port, an output port, a processor and a memory including program code executable by the processor. The program code includes code for receiving, at the input port, a data frame including a plurality of data elements. The program code further includes code for transforming the data frame into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements. Also included within the program code is code for generating a modulated signal in accordance with the transformed data elements of the transformed data matrix.
In additional aspect, the disclosure relates to a nontransitory computer readable medium including program instructions for execution by a processor in a signal transmitter. The program instructions include instructions for causing the processor to receive, at an input port of the signal transmitter, a data frame including a plurality of data elements. The program instructions further cause the processor to transform the data frame into a transformed data matrix having a plurality of transformed data elements wherein ones of the plurality of transformed data elements are based upon multiple ones of the plurality of data elements. In addition, the program instructions cause the processor to generate a modulated signal in accordance with the transformed data elements of the transformed data matrix.
In yet a further aspect the disclosure pertains to a nontransitory computer readable medium including program instructions for execution by a processor in a signal transmitter. The program instructions include instructions for causing the processor to receive, at an input port of the signal transmitter, a data frame including a plurality of data elements. The instructions also cause the processor to transform the data frame into a transformed data matrix having a plurality of transformed data elements wherein each of the plurality of transformed data elements is based upon each of the plurality of data elements. In addition, the program instructions cause the processor to generate a modulated signal in accordance with the transformed data elements of the transformed data matrix.
In another aspect, the present disclosure describes a method of providing a modulated signal useable in a signal transmission system. The method of this aspect includes establishing an original data frame having a first dimension of at least N elements and a second dimension of at least N elements, wherein N is greater than one, transforming the original data frame in accordance with a timefrequency transformation so as to provide a transformed data matrix, and generating the modulated signal in accordance with elements of the transformed data matrix.
In another aspect, the present disclosure describes a method for modulating data for transmission within a communication system. The method of this aspect includes establishing a timefrequency shifting matrix of dimension N×N, wherein N is greater than one, combining the timefrequency shifting matrix with a data frame to provide an intermediate data frame, providing a transformed data matrix by permuting elements of the intermediate data frame, and generating a modulated signal in accordance with elements of the transformed data matrix.
In another aspect, the present disclosure describes a method of data modulation. The method of this aspect includes arranging a set of data elements into an original data frame having a first dimension of N elements and a second dimension of N elements, where N is greater than one, transforming the original data frame in accordance with a timefrequency shifting matrix so as to form an intermediate data matrix having at least N^{2 }elements, providing a transformed data matrix by permuting at least a portion of the elements of the intermediate data matrix, and generating a modulated signal based upon elements of the transformed data matrix.
In another aspect, the present disclosure describes a method of receiving data. The method of this aspect includes receiving data signals corresponding to a transmitted data frame comprised of a set of data elements, constructing, based upon the data signals, a received data frame having a first dimension of at least N elements and a second dimension of at least N elements, where N is greater than one, inverse permuting at least a portion of the elements of the received data frame so as to form an nonpermuted data frame, and inverse transforming the nonpermuted data frame in accordance with first inversetransformation matrix so as to form a recovered data frame corresponding to a reconstructed version of the transmitted data frame.
In another aspect, the present disclosure describes a method of data transmission. The method of this aspect includes arranging a set of data elements into an original data frame having a first dimension of N elements and a second dimension of N elements, where N is greater than one, and transforming the original data frame in accordance with a transformation matrix to form a first transformed data matrix having at least N^{2 }transformed data elements wherein each of the transformed data elements is based upon a plurality of the data elements of the original data frame and wherein a first dimension of the first transformed data matrix corresponds to a frequency shift axis and a second dimension corresponds to a time shift axis. The method of this aspect further includes forming a permuted data matrix by permuting at least a portion of the elements of the first transformed data matrix so as to shift the at least a portion of the elements with respect to the time shift axis, transforming the permuted data matrix using a frequencyshift encoding matrix to form a transmit frame, and generating a modulated signal in accordance with elements of the transmit frame.
In another aspect, the present disclosure describes a method of receiving data. The method of this aspect includes receiving, on one or more carrier waveforms, signals representing a plurality of data elements of an original data frame wherein each of the data elements are represented by cyclically time shifted and cyclically frequency shifted versions of a known set of waveforms, generating, based upon the signals, a received data frame having a first dimension of at least N elements and a second dimension of at least N elements, where N is greater than one, wherein the first dimension corresponds to a frequency shift axis and the second dimension corresponds to a time shift axis. The method of this aspect further includes performing, using a decoding matrix, an inverse transformation operation with respect to elements of the received data frame so as to yield a nontransformed matrix, and generating, based upon the nontransformed matrix, a recovered data frame comprising an estimate of the original data frame.
In another aspect, the present disclosure describes a new signal modulation technique which involves spreading data symbols over a large range of times, frequencies, and/or spectral shapes (waveforms). Such method, termed “Orthonormal TimeFrequency Shifting and Spectral Shaping (“OTFSSS”) when spectral shaping is employed or, more generally, “OTFS”, operates by sending data in what are generally substantially larger “chunks” or frames than the data frames used in prior methods. That is, while prior art methods might encode and send units or frames of “N” symbols over a communications link over a particular time interval, OTFS contemplates that frames of N^{2 }symbols are transmitted (often over a relatively longer time interval). With OTFS modulation, each data symbol or element that is transmitted is extensively spread in a novel manner in time, frequency and/or spectral shape space. At the receiving end of the connection, each data symbol is resolved based upon substantially the entire frame of N^{2 }received symbols.
In another aspect, disclosed herein is a wireless communication method predicated upon spreading input data over time, frequency and potentially spectral shape using convolution unit matrices (data frames) of N×N (N^{2}). In general, either all N^{2 }of the data symbols are received over N particular spreading time intervals (each composed of N time slices), or no such symbols are received. During the transmission process, each N×N data frame matrix will typically be multiplied by a first N×N timefrequency shifting matrix, permuted, and then multiplied by a second N×N spectralshaping matrix, thereby mixing each data symbol across the entire resulting N×N matrix (which may be referred to as the TFSSS data matrix). Columns from the TFSSS data matrix are then selected, modulated, and transmitted, on a one element per time slice basis. At the receiver, a replica TFSSS matrix is reconstructed and deconvolved, yielding a reconstruction of the input data.
Embodiments of the systems and methods described herein may use novel timefrequency shifting and spectral shaping codes to spread data across time, spectrum, waveform, and/or spectralshape. In such embodiments timeshifting techniques, frequencyshifting techniques and, optionally, spectralshaping techniques may be used in conjunction to transmit data at high rates in a manner unusually resistant to problems caused by Doppler shifts, multipath effects, and background noise.
During the signal transmission process, an OTFS transmitter may subdivide and transmit each data element or symbol over a cyclically varying range of frequencies and over a series of spreading time intervals. Often this will require that each data element or symbol be transmitted over a somewhat longer period of time than is utilized for transmission data frames in other communication systems. Notwithstanding these potentially longer transmission periods, the OTFS system is capable of achieving superior data rate performance by using a complex multiplexing methods premised on the convolution and deconvolution schemes discussed herein. Through use of such methods a comparatively large amount of information may be included within each transmitted signal. In particular, the relatively large number (i.e., N^{2}) of data symbols or elements transmitted during each data frame using the convolution and deconvolution schemes disclosed herein enable comparatively high data rates to be attained despite the diminution in data rate which could otherwise result from division of a single data element or symbol over N timespreading intervals. Moreover, because each data symbol is typically subdivided and sent over a plurality of signals, signal processing schemes described herein may be employed to permit data symbols to be recovered even in the event of loss of one or more of the plurality of transmitted signals. In addition, such schemes may be employed to compensate for losses due to common wireless communications link impairments, such as Doppler shift and multipath effects.
For example, whereas with prior art, if by chance Doppler effects caused by one wireless signal from a first transmitter fall on the same frequency as another signal from a first or second transmitter (for multipath effects, a signal from a moving first or second transmitter that hits an object at an arbitrary angle to the receiver can produce a Doppler distorted reflection or echo signal of the first or second transmitter that also reaches the receiver), this could result in confusion, ambiguity, and data loss. By contrast, by cyclically shifting the frequency and sending an element of data over a plurality of time intervals, the impact of a Doppler “collision” is substantially minimized—at most there will be a brief transient effect resulting in the loss of only one of a plurality of signals used to transmit a particular data symbol or element. The effects of other communications link impairments, such as multipath effects, can also be minimized because the cyclically shifting frequency provides yet another way to compensate for multipath effects.
There exist at least two ways in which a data element or symbol may be partitioned across a time range of cyclically shifting frequencies, and thus two basic forms of the OTFS method. In a first form of the OTFS method, the data from a single symbol is convolved and partitioned across multiple time slices, and ultimately transmitted as a series of time slices, on a per time slice basis. When this transmission scheme is used the cyclicallyshifting frequency is accomplished over a plurality of time spreading intervals. Thus, for this first form of the OTFS method, the basic unit of data transmission operates on a time slice basis.
In a second form of the OTFS method, the data is ultimately transmitted as a series of waveforms with characteristic frequencies, where each waveform lasts for a spreading interval of time generally consisting of N time slices. When this transmission scheme is used the cyclicallyshifting frequency is accomplished over a plurality of time spreading intervals. Thus, for this second form of the OTFS method, the basic unit of data transmission operates over a relatively longer spreading time interval comprised of N time slices. Unless otherwise specified, the discussion within the remainder of this disclosure will focus on the first form of the OTFS method in which the basic unit of data transmission operates on a time slice basis.
Again considering the first form of the OTFS method, in one embodiment this form of the OTFS method contemplates formation of an N×N data frame matrix having N^{2 }symbols or elements and multiplication of this data frame by a first N×N timefrequency shifting matrix. The result of this multiplication is optionally permuted and, following permutation, optionally multiplied by a second N×N spectral shaping matrix. As a result, the N^{2 }data elements in the frame of data are essentially mixed or distributed throughout the resulting N×N matrix product, here called a “Time Frequency Shifted” data matrix or “TFS” data matrix. If the optional spectral shaping is used, the resulting N×N matrix product may be referred to as a “Time Frequency Shifted and Spectral Shaped” data matrix or “TFSSS” data matrix. Thus, for example, a single symbol or element in row 1 column 1 of the N×N frame of data may end up being distributed over all rows and columns of the resulting N×N TFS or TFSSS data matrix (in what follows the term TFS may connote either the TFS or TFSSS data matrix).
The contents (i.e. the individual elements) of this TFS data matrix may then be selected, modulated, and transmitted. Usually N elements at a time from this TFS data matrix (often a column from the TFS data matrix) are selected to be sent over one spreading interval of time, thus often requiring N spreading intervals of time to transmit the entire contents of the TFS data matrix. This spreading interval of time in turn is usually composed of at least N time slices. During each time slice, one element from the most recent selection of N elements (for example, from the selected column of the TFS data matrix) is selected, modulated, and transmitted.
At the receiving end, the process operates generally in reverse. The individual elements of the TFS data matrix are received over various time slices and various time spreading intervals, allowing the receiver to reassemble a replica (which may not be a perfect replica due to communications link impairment effects) of the original TFS data matrix. Using its knowledge of the first N×N timefrequency shifting matrix, the optional permutation process, the second N×N spectral shaping matrix, and the selection process used to select different elements of the TFS data matrix, as well as various noise reduction or compensation techniques to overcome impairment effects, the receiver will then reconstruct the original N×N data frame matrix of N^{2 }symbols or elements. Because each data symbol or element from the original data frame is often spread throughout the TFS data matrix, often most or the entire TFS matrix will need to be reconstructed in order to solve for the original data symbol or element. However, by using noise reduction and compensation techniques, minor data losses during transmission can often be compensated for.
In some embodiments, advanced signal modulation schemes utilizing cyclically time shifted and cyclically frequency shifted waveforms may be utilized to correct channel impairments in a broad range of situations. For example, in one aspect the OTFS method may contemplate transfer of a plurality of data symbols using a signal modulated in a manner which effectively compensates for the adverse effects of echo reflections and frequency offsets. This method will generally comprise distributing this plurality of data symbols into one or more N×N symbol matrices and using these one or more N×N symbol matrices to control the signal modulation occurring within a transmitter. Specifically, during the transmission process each data symbol within an N×N symbol matrix is used to weight N waveforms. These N waveforms are selected from an N^{2 }sized set of all permutations of N cyclically time shifted and N cyclically frequency shifted waveforms determined according to an encoding matrix U. The net result produces, for each data symbol, N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms. Generally this encoding matrix U is chosen to be an N×N unitary matrix that has a corresponding inverse decoding matrix U^{H}. Imposition of this constraint means that the encoding matrix U produces results which can generally be decoded.
Continuing with this example, for each data symbol in the N×N symbol matrix, the transmitter may sum the corresponding N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms, and by the time that the entire N×N symbol matrix is so encoded, produce N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms. The transmitter will then transmit these N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms, structured as N composite waveforms, over any combination of N time blocks or frequency blocks.
To receive and decode this transmission, the transmitted N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms are subsequently received by a receiver which is controlled by the corresponding decoding matrix U^{H}. The receiver will then use this decoding matrix U^{H }to reconstruct the original symbols in the various N×N symbol matrices.
This process of transmission and reception will normally be performed by various electronic devices, such as a microprocessor equipped, digital signal processor equipped, or other electronic circuit that controls the convolution and modulation parts of the signal transmitter. Similarly the process of receiving and demodulation will also generally rely upon a microprocessor equipped, digital signal processor equipped, or other electronic circuit that controls the demodulation, accumulation, and deconvolution parts of the signal receiver. However, although the exemplary techniques and systems disclosed herein will often be discussed within the context of a wireless communication system comprised of at least one wireless transmitter and receiver, it should be understood that these examples are not intended to be limiting. In alternative embodiments, the transmitter and receiver may be optical/opticalfiber transmitters and receivers, electronic wire or cable transmitters and receivers, or other types of transmitters in receivers. In principle, more exotic signal transmission media, such as acoustic signals and the like, may also be utilized in connection with the present methods.
As previously discussed, regardless of the media (e.g. optical, electrical signals, or wireless signals) used to transmit the various waveforms, these waveforms can be distorted or impaired by various signal impairments such as various echo reflections and frequency shifts. As a result, the receiver will often receive a distorted form of the original signal. Here, embodiments of the OTFS method make use of the insight that cyclically time shifted and cyclically frequency shifted waveforms are particularly useful for detecting and correcting for such distortions.
Because communications signals propagate through their respective communications media at a finite speed (often at or near the speed of light), and because the distance from a transmitter to a receiver is usually substantially different than the distance between the transmitter, the place(s) where the echo is generated, and the distance between the place(s) where the echo is generated and the receiver, the net effect of echo reflections is that both an originally transmitted waveform and timeshifted versions thereof are received at the receiver, thereby resulting in a distorted composite signal. However, embodiments of the OTFS method utilizing cyclically time shifted waveforms may be employed to counteract this distortion. In particular, a time deconvolution device at the receiver may operate to analyze the cyclically time varying patterns of these waveforms, determine the repeating patterns, and use these repeating patterns to help decompose the echo distorted signal back into various timeshifted versions of the various signals. The time deconvolution device can also determine how much of a timeoffset (or multiple time offsets) is or are required to enable the time delayed echo signal(s) to match up with the originally transmitted signal. This time offset value, which may be referred to herein as a time deconvolution parameter, can both give useful information as to the relative position of the echo location(s) relative to the transmitter and receiver, and can also help the system characterize some of the signal impairments that occur between the transmitter and receiver. This can help the communications system automatically optimize itself for better performance.
In addition to echo reflections, other signal distortions may occur that can result in one or more frequency shifts. For example, a Doppler shift or Doppler effects can occur when a wireless mobile transmitter moves towards or away from a stationary receiver. If the wireless mobile transmitter is moving towards the stationary receiver, the wireless waveforms that it transmits will be offset to higher frequencies, which can cause confusion if the receiver is expecting signals modulated at a lower frequency. An even more confusing result can occur if the wireless mobile transmitter is moving perpendicular to the receiver, and there is also an echo source (such as a building) in the path of the wireless mobile transmitter. Due to Doppler effects, the echo source receives a blue shifted (higher frequency) version of the original signal, and reflects this blue shifted (higher frequency) version of the original signal to the receiver. As a result, the receiver will receive both the originally transmitted “direct” wireless waveforms at the original lower frequency, and also a timedelayed higher frequency version of the original wireless waveforms, causing considerable confusion.
It has been found that the use of cyclically time shifted waveforms and cyclically frequency shifted waveforms may help address this type of problem. In particular, it has been found that the cyclic variation yields important pattern matching information which may enable the receiver to determine which portions of a received signal have been distorted and the extent of such distortion. In one embodiment these cyclically varying signals allow the receiver to perform a twodimensional (e.g. time and frequency) deconvolution of the received signal. For example, the frequency deconvolution portion of the receiver can analyze the cyclically frequency varying patterns of the waveforms, essentially do frequency pattern matching, and decompose the distorted signal into various frequency shifted versions of the various signals. At the same time, this portion of the receiver can also determine how much of a frequency offset is required to cause the frequency distorted signal to match up with the originally transmitted signal. This frequency offset value, herein termed a “frequency deconvolution parameter”, can give useful information as to the transmitter'"'"'s velocity relative to the receiver. This may facilitate characterization of some of the frequency shift signal impairments that occur between the transmitter and receiver.
As before, the time deconvolution portion of the receiver can analyze the cyclically time varying patterns of the waveforms, again do time pattern matching, and decompose the echo distorted signal back into various timeshifted versions of the original signal. The time deconvolution portion of the receiver can also determine how much of a timeoffset is required to cause the time delayed echo signal to match up with the original or direct signal. This time offset value, again called a “time deconvolution parameter”, can also give useful information as to the relative positions of the echo location(s), and can also help the system characterize some of the signal impairments that occur between the transmitter and receiver.
The net effect of both the time and frequency deconvolution, when applied to transmitters, receivers, and echo sources that potentially exist at different distances and velocities relative to each other, is to allow the receiver to properly interpret the impaired echo and frequency shifted communications signals.
Further, even if, at the receiver, the energy received from the undistorted form of the originally transmitted signal is so low as to have a undesirable signal to noise ratio, by applying the appropriate time and frequency offsets or deconvolution parameters, the energy from the time and/or frequency shifted versions of the signals (which would otherwise be contributing to noise) can instead be harnessed to contribute to the signal instead.
As before, the time and frequency deconvolution parameters can also provide useful information as to the relative positions and velocities of the echo location(s) relative to the transmitter and receiver, as well as the various velocities between the transmitter and receiver. These in turn can help the system characterize some of the signal impairments that occur between the transmitter and receiver, as well as assist in automatic system optimization methods.
Thus in some embodiments, the OTFS system may also provide a method for an improved communication signal receiver where, due to either one or the combination of echo reflections and frequency offsets, multiple signals due to echo reflections and frequency offsets result in the receiver receiving a time and/or frequency convolved signal representing time and/or frequency shifted versions of the N^{2 }summationsymbolweighed cyclically time shifted and frequency shifted waveforms previously sent by the transmitter. Here, the improved receiver will further perform a time and/or frequency deconvolution of the impaired signal to correct for various echo reflections and frequency offsets. This improved receiver method will result in both time and frequency deconvolved results (i.e. signals with higher quality and lower signal to noise ratios), as well as various time and frequency deconvolution parameters that, in addition to automatic communications channel optimization, are also useful for other purposes as well. These other purposes can include channel sounding (i.e. better characterizing the various communication system signal impairments), adaptively selecting modulation methods according to the various signal impairments, and even improvements in radar systems.
For a better understanding of the nature and objects of various embodiments of the invention, reference should be made to the following detailed description taken in conjunction with the accompanying drawings, wherein:
One unique aspect of the signal modulation techniques described herein is the concept of spreading the data of a single symbol over a relatively large range of times, frequencies, and spectral shapes. In contrast, prior communication systems have been predicated upon assigning a given data symbol to a specific timespreading interval or time slice uniquely associated with such data symbol. As is discussed below, the disclosed OTFS method is based at least in part upon the realization that in many cases various advantages may accrue from spreading the data of a single symbol over multiple timespreading intervals shared with other symbols. In contrast with prior art modulation techniques, the OTFS method may involve convolving a single data symbol over both a plurality of time slots, a plurality of frequencies or spectral regions (spread spectrum), and a plurality of spectral shapes. As is described below, this approach to data convolution results in superior performance over impaired communications links.
Turning now to
Unfortunately, use of a onedimensional channel model presents a number of fundamental problems. First, the onedimensional channel models employed in existing communication systems are nonstationary; that is, the symboldistorting influence of the communication channel changes from symbol to symbol. In addition, when a channel is modeled in only one dimension it is likely and possible that certain received symbols will be significantly lower in energy than others due to “channel fading”. Finally, onedimensional channel state information (CSI) appears random and much of it is estimated by interpolating between channel measurements taken at specific points, thus rendering the information inherently inaccurate. These problems are only exacerbated in multiantenna (MIMO) communication systems. As is discussed below, embodiments of the OTFS method described herein can be used to substantially overcome the fundamental problems arising from use of a onedimensional channel model.
As is indicated below by Equation (1), in one aspect the OTFS method recognizes that a wireless channel may be represented as a weighted superposition of combination of time and Doppler shifts:
In contrast to the parameters associated with existing channel models, the timefrequency weights (τ,) of Equation (1) are twodimensional and are believed to fully characterize the wireless channel. The timefrequency weights (τ,) are intended to represent essentially all of the diversity branches existing in the wireless channel. This is believed to substantially minimize the fading effects experienced by the OTFS system and other communication systems generally based upon twodimensional channel models relative to the fading common in systems predicated upon onedimensional models. Finally, in contrast to the nonstationary, onedimensional channel models employed in conventional communication systems, the timefrequency weights (τ,) of Equation (1) are substantially stationary; that is, the weights change very slowly relative to the time scale of exemplary embodiments of the OTFS system.
Use of the twodimensional channel model of Equation (1) in embodiments of the OTFS communication system affords a number of advantages. For example, use of the channel model of Equation (1) enables both channel multipath delay and Doppler shift to be accurately profiled simultaneously. Use of this model and the OTFS modulation techniques described herein also facilitate the coherent assembly of channel echoes and the minimization of fading phenomena, since every symbol experience substantially all of the diversity branches present within the channel. Given that the twodimensional channel model is essentially stationary, every symbol is deterministically distorted (smeared) according to substantially the same twodimensional pattern. This stable, accurate characterization of the communication channel in two dimensions on an ongoing basis further enables the OTFS system to minimize data distortion by “customizing” how each bit is delivered across the channel. Finally, use of a twodimensional channel model enables effective signal separation by decoupling and eliminating mutual interference between multiple sources.
Attention is now directed to
The preequalizer 210 is used to model a predistortion transfer function h_{t }that can be used to make up for changing channel conditions in the channel model h_{c }based on feedback received over the channel from the receive side of the model, as determined by measurements made by the receiver/demodulator 260 and/or the post equalizer 270. The transmitter/modulator 220 uses modulation schemes described herein to transmit the data over the channel 230.
The receiver/demodulator 260 demodulates the signal received over the channel 230. The received signal has been distorted by time/frequency selective fading, as determined by the channel transfer function h_{c}, and includes the additive noise 240. The receiver/demodulator 260 and the post equalizer 270 utilize methods discussed herein to reduce the distortion caused by the time/frequency selective fading and additive noise due to the channel conditions. The mathematical model 200 can be used to determine the nature of the equalized data D_{eq }by performing a mathematical combination of three transfer functions operating on the original data D. The three transfer functions include the transmitter transfer function h_{t}, the channel transfer function h_{c }and the equalizer transfer function h_{r}.
Embodiments of the OTFS methods and systems described herein are based, in part, upon the realization that spreading the data for any given symbol over time, spectrum, and/or spectral shapes in the manner described herein yields modulated signals which are substantially resistant to interference, particularly interference caused by Doppler effects and multipath effects, as well as general background noise effects. Moreover, the OTFS method is believed to require less precise frequency synchronization between receiver and transmitter than is required by existing communication systems (e.g., OFDM systems).
In essence, the OTFS method convolves the data for a group of N^{2 }symbols (herein called a “frame”) over both time, frequency, and in some embodiments spectral shape in a way that results in the data for the group of symbols being sent over a generally longer period of time than in prior art methods. Use of the OTFS method also results in the data for any given group of symbols being accumulated over a generally longer period of time than in prior art methods. However, in certain embodiments the OTFS method may nonetheless enable favorable data rates to be achieved despite the use of such longer transmission periods by exploiting other transmission efficiencies enabled by the method. For example, in one embodiment a group of symbols may be transmitted using the same spreadspectrum code. Although this could otherwise result in confusion and ambiguity (since each symbol would not be uniquely associated with a code), use of the OTFS method may, for example, enable the symbols to be sent using different (but previously defined) spreadspectrum convolution methods across a range of time and frequency periods. As a consequence, when all of the data corresponding to the symbols is finally accumulated within the receiver, the entire frame or group of symbols may be reconstructed in a manner not contemplated by prior art techniques. In general, one tradeoff associated with the disclosed approach is that either an entire multisymbol frame of data will be correctly received, or none of the frame will be correctly received; that is, if there is too much interference within the communication channel, then the ability to successfully deconvolve and retrieve multiple symbols may fail. However, as will be discussed, various aspects of the OTFS may mitigate any degradation in performance which would otherwise result from this apparent tradeoff.
In one aspect a method of OTFS communication involves transmitting at least one frame of data ([D]) from the transmitting device 310 to the receiving device 330 through the communication channel 320, such frame of data comprising a matrix of up to N^{2 }data elements, N being greater than 1. The method comprises convolving, within the OTFS transceiver 3151, the data elements of the data frame so that the value of each data element, when transmitted, is spread over a plurality of wireless waveforms, each waveform having a characteristic frequency, and each waveform carrying the convolved results from a plurality of said data elements from the data frame [D]. Further, during the transmission process, cyclically shifting the frequency of this plurality of wireless waveforms over a plurality of times so that the value of each data element is transmitted as a plurality of cyclically frequency shifted waveforms sent over a plurality of times. At the receiving device 330, the OTFS transceiver 3152 receives and deconvolves these wireless waveforms thereby reconstructing a replica of said at least one frame of data [D]. In the exemplary embodiment the convolution process is such that an arbitrary data element of an arbitrary frame of data ([D]) cannot be guaranteed to be reconstructed with full accuracy until substantially all of these wireless waveforms have been transmitted and received.
Attention is now directed to
xεC
^{N×N }
As is illustrated in
φ_{t}=M(x)=Σx(i,j)φ_{i,j}φ_{i,j}⊥φ_{k,l }
Referring to
φ_{r}y=D(φ_{r})
In one embodiment the OTFS transceiver 4500 may be characterized by a number of variable parameters including, for example, delay resolution (i.e., digital time “tick” or clock increment), Doppler resolution, processing gain factor (block size) and orthonormal basis function. Each of these variable parameters may be represented as follows.
Delay resolution (digital time tick):
Doppler resolution:
Processing gain factor (block size):
N>0
Orthonormal basis of C^{N×1 }(spectral shapes):
U={u_{1},u_{2}, . . . ,u_{N}}
As is illustrated by
where b_{1}, b_{2 }. . . b_{N }are illustrated in
Π(h*x)=Π(h)·Π(x)
in particular:
Π(δ_{(τ,o)}*x)=L_{τ}·Π(x)
Π(δ_{0,w})*x)=M_{w}·Π(x)
The Heisenberg representation provides that:
where L_{t }and M_{w }are respectively representative of cyclic time and frequency shifts and may be represented as:
The demodulator 4520 takes a received waveform and transforms it into a TF matrix yεC^{N×N }defined in terms of the Wigner transform and the spectral shapes:
Main property of M and D (Stone von Neumann theorem):
D(h°M(x))=h*x
where:
h(τ,w)≈a(τΔT,wΔF)
As illustrated in
y
{circumflex over (x)}
Throughout this description, the use of matrix terminology should be understood as being a concise description of the various operations that will be carried out by either the OTFS transceiver 3151 or the OTFS transceiver 3152. Thus the series of steps used to obtain the coefficients of a particular matrix generally correspond to a set of instructions for the transmitter or receiver electronic circuitry (e.g., the various components of the transmitter 405 and the receiver 455 illustrated in
Thus, when the discussion speaks of multiplying matrices, each data element in the matrix formed by the multiplication can be understood in terms of various multistep operations to be carried out by the transmitter or receiver electronic circuitry (e.g., the transmitter 405 or the receiver 455 as illustrated in
Put into matrix terminology, the OTFS method of convolving the data for a group of symbols over both time, spectrum, and tone or spectralshape can be viewed as transforming the data frame with N^{2 }information elements (symbols) to another new matrix with N^{2 }elements whereby each element in the new transformed matrix, (here called the TFS data matrix) carries information about all elements of the original data frame. In other words the new transformed TFS data matrix will generally carry a weighted contribution from each element of the original data frame matrix [D]. Elements of this TFS data matrix are in turn transmitted and received over successive time intervals.
As previously discussed, in embodiments of the OTFS method the basic unit of convolution and deconvolution (convolution unit) is composed of a matrix of N^{2 }symbols or data elements. Over each time interval, a different waveform may be used for each data element. By contrast, prior art methods generally use the same waveform for each data element. For consistency, the N^{2 }units of data will generally be referred to in this specification as a “data frame”. N may be any value greater than one, and in some embodiments will range from 64 to 256.
One distinction between the OTFS method and conventional modulation schemes may be appreciated by observing that a basic unit of convolution, transmission, reception and deconvolution for a prior art communications protocol may be characterized as a data frame of n symbols or elements “d” operated on spreading codes that send the data for n symbols over one spreading interval time where:
[D_{1xn}]=[d_{1}d_{2 }. . . d_{n}]
In contrast, embodiments of the OTFS method generally use a different basic unit of convolution, transmission, reception, and deconvolution. In particular, such OTFS embodiments will typically use a larger data frame [D_{N×N}] composed of N^{2 }elements or symbols “d” that, as will be discussed, send the data for these N^{2 }elements over a plurality of spreading interval times (often the plurality is N). The data frame [D_{N×N}] may be expressed as:
In general, references herein to a frame of data may be considered to be a reference to the N×N or N^{2 }matrix such as the one shown above, where at least some elements in the matrix may be zero or null elements. In some embodiments, a frame of data could be nonsquare, or N×M, where N # M.
As previously discussed, the OTFS method will spread this group of N^{2 }symbols across a communications link over multiple spreading intervals of time (usually at least N spreading intervals or times), where each individual spreading interval of time is composed of at least N time slices. Note that due to potential overhead for synchronization and identification purposes, in some embodiments, extra time slices and/or extra spreading intervals of time may be allocated to provide room for this overhead. Although for clarity of presentation this overhead will generally be ignored, it should be understood that the disclosure is intended to also encompass methods in which such overhead exists.
In exemplary embodiments of the OTFS method the data will thus be transmitted as a complex series of waveforms, usually over wireless radio signals with frequencies usually above 100 MHz, and often above 1 GHz or more. These radio frequencies are then usually received over at least N spreading time intervals, where each spreading time interval is often composed of at least N timeslices. Once received, the original data frame will be deconvolved (i.e. solved for) and the most likely coefficients of the original group of symbols are reconstructed. It should be evident that in order to successfully deconvolve or solve for the original data frame, the receiver will usually have prior knowledge of the time, spectrum, and tone or spectralshape spreading algorithms used by the transmitter.
Attention is now directed to
 1: Construct the matrix product of a first N×N matrix [U1] and [D] (often written as either [U_{1}]*[D] or more simply [U_{1}][D]—here both the “*” and simple close association (e.g. [U_{1}][D]) both are intended to represent matrix multiplication)(stage 342).
 2: Optionally permute [U_{1}][D] by a permutation operation P in order to create a new N×N matrix (stage 344). In general, any invertible permutation operation may be used. P may be an identity operation, or alternatively may be a permutation operation that essentially translates the columns of the original N×N [U_{1}][D] matrix to diagonal elements of a transformed [U_{1}][D]′ matrix.
 3: Upon completing the permutation, optionally multiply the permuted result by a second N×N [U_{2}] matrix (for spectral shaping for example), forming:
 [P([U_{1}][D])][U_{2}] (stage 348).
 4: Transmit this signal, according to methods discussed below (stage 350).
 In one embodiment the permutation operation P may optionally be of the form:
b
_{i,j}
=a
_{i,(j−i)mod N }
where [a] is the original matrix (here [U_{1}][D]), and [b] is the new matrix (here P([U_{1}][D]).
For sake of simplicity, the result of this permutation operation may be written as P([U_{1}][D]).
b
_{i,j}
=a
_{i,(j+i)mod N }
Yet another permutation option is illustrated in
Here [U_{1}] and [U_{2}], if being used, can both be unitary N×N matrices, usually chosen to mitigate certain impairments on the (often wireless) communications link, such as wide band noise, narrowband interference, impulse noise, Doppler shift, crosstalk, etc. To do this, rather than simply selecting [U_{1}] and [U_{2}] to be relatively trivial identity matrices [I], or matrices with most of the coefficients simply being placed along the central diagonal of the matrix, [U_{1}] and [U_{2}] will usually be chosen with nonzero coefficients generally throughout the matrix so as to accomplish the desired spreading or convolution of the convolution unit [D] across spectrum and tone or spectralshape space in a relatively efficient and uniform manner. Usually, the matrix coefficients will also be chosen to maintain orthogonality or to provide an ability to distinguish between the different encoding schemes embodied in the different rows of the respective matrices, as well as to minimize autocorrelation effects that can occur when radio signals are subjected to multipath effects.
In reference to the specific case where [U_{1}] may have rows that correspond to pseudorandom sequences, it may be useful to employ a permutation scheme where each successive row in the matrix is a cyclically rotated version of the pseudorandom sequence in the row above it. Thus the entire N×N matrix may consist of successive cyclically rotated versions of a single pseudorandom sequence of length N.
In principle, [U_{1}] and [U_{2}], if both are being used, may be a wide variety of different unitary matrices. For example, [U_{1}] may be a Discrete Fourier Transform (DFT) matrix and [U_{2}] may be a Hadamard matrix. Alternatively [U_{1}] may be a DFT matrix and [U_{2}] may be a chirp matrix. Alternatively [U_{1}] may be a DFT matrix and [U_{2}] may also be a DFT matrix, and so on. Thus although, for purposes of explaining certain aspects of the OTFS method, certain specific examples and embodiments of [U_{1}] and [U_{2}] will be given, these specific examples and embodiments are not intended to be limiting.
Note that a chirp matrix, [V], is commonly defined in signal processing as a matrix where, if Ψ the chirp rate,
 [V]=diag(Ψ, Ψ^{2}, . . . Ψ^{n}), Ψ=e^{iψ}, and frequency=e^{iω} where ω is the initial center frequency of the spectrum.
Alternatively, a different chirp matrix may be used, filled with elements of the form:
Where j is the matrix row, k is the matrix column, and N is the size of the matrix.
Other commonly used orthonormal matrices, which may be used for [U_{1}], or [U_{2}] or [U_{3}] (to be discussed), include Discrete Fourier matrices, Polynomial exponent matrices, harmonic oscillatory, matrices, the previously discussed Hadamard matrices, Walsh matrices, Haar matrices, Paley matrices, Williamson matrices, Msequence matrices, Legendre matrices, Jacobi matrices, Householder matrices, Rotation matrices, and Permutation matrices. The inverses of these matrices may also be used.
As will be discussed, in some embodiments, [U_{1}] can be understood as being a timefrequency shifting matrix, and [U_{2}] can be understood as being a spectral shaping matrix. In order to preserve readability, [U_{1}] will thus often be referred to as a first timefrequency shifting matrix, and [U_{2}] will thus often be referred to as a second spectral shaping matrix. However use of this nomenclature is also not intended to be limiting. In embodiments in which the optional permutation or multiplication by a second matrix [U2] is not performed, the [U1] matrix facilitates time shifting by providing a framework through which the elements of the resulting transformed data matrix to be transmitted at different times (e.g., on a column by column basis or any other ordered basis).
Turning to some more specific embodiments, in some embodiments, [U_{1}] may have rows that correspond to Legendre symbols, or spreading sequences, where each successive row in the matrix may be a cyclically shifted version of the Legendre symbols in the row above it. These Legendre symbols will occasionally also be referred to in the alternative as base vectors, and occasionally as spectrumspreading codes.
In some embodiments, [U_{2}] may chosen to be a Discrete Fourier transform (DFT) matrix or an Inverse Discrete Fourier Transform matrix (IDFT). This DFT and IDFT matrix can be used to take a sequence of real or complex numbers, such as the N×N (P[U_{1}][D]) matrix, and further modulate P([U_{1}][D]) into a series of spectral shapes suitable for wireless transmission.
The individual rows for the DFT and IDFT matrix [U_{2}] will occasionally be referred in the alternative as Fourier Vectors. In general, the Fourier vectors may create complex sinusoidal waveforms (tone or spectralshapes) of the type:
where, for an N×N DFT matrix, X is the coefficient of the Fourier vector in row k, column N of the DFT matrix, and j is the column number. The products of this Fourier vector can be considered to be tones or spectralshapes.
Although certain specific [U_{1}] and [U_{2}] can be used to transmit any given data frame [D], when multiple data frames [D] are being transmitted simultaneously, the specific [U_{1}] and [U2] chosen may vary between data frames [D], and indeed may be dynamically optimized to avoid certain communications link impairments over the course of transmitting many data frames [D] over a communications session.
This process of convolution and modulation will normally be done by an electronic device, such as a microprocessor equipped, digital signal processor equipped, or other electronic circuit that controls the convolution and modulation parts of the wireless radio transmitter. Similarly the process of receiving and demodulation will also generally rely upon a microprocessor equipped, digital signal processor equipped, or other electronic circuit that controls the demodulation, accumulation, and deconvolution parts of the wireless radio receiver.
Thus again using matrix multiplication, and again remembering that these are all N×N matrices, [P([U_{1}][D])][U_{2}], where [U_{2}] is optional, represents the TFS data matrix that the transmitter will distribute over a plurality of time spreading intervals, time slices, frequencies, and spectral shapes. Note again that as a result of the various matrix operation and optional permutation steps, a single element or symbol from the original N×N data matrix [D] after modulation and transmission, will be distributed throughout the different time spreading intervals, time slices, frequencies, and spectral shapes, and then reassembled by the receiver and deconvolved back to the original single data element of symbol.
In the example of
The matrix [U_{2}] 604 may be a DFT or IDFT matrix and is often designed to spectrally shape the signals. For example, in some embodiments the matrix [U_{2}] 604 may contain the coefficients to direct the transmitter circuits of the OTFS modulator 430 to transform the signals over time in a OFDM manner, such as by quadratureamplitude modulation (QAM) or phaseshift keying, or other scheme.
Usually the matrix [D] 601 will be matrix multiplied by the matrix [U_{1}] 602 by the digital encoder 665 at stage 610, and the matrix product of this operation [U_{1}][D] then optionally permuted by the digital encoder 665 forming P([U_{1}][D]) (stage 611). In embodiments in which a spectral shaping matrix is utilized, the digital encoder 665 multiplies matrix [U_{1}][D] by matrix [U_{2}] 604 forming an N×N TFS data matrix, which may also be referred to herein as an OFTS transmission matrix (stage 614).
The various elements of the TFS matrix are then selected by the OTFS analog modulator 670, usually a column of N elements at a time, on a single element at a time basis (stage 616) The selected elements are then used to generate a modulated signal that is transmitted via an antenna 680 (stage 618). More specifically, in one embodiment the particular real and imaginary components of each individual TFS matrix element are used to control a time variant radio signal 620 during each time slice. Thus one Nelement column of the TFS matrix will usually be sent during each timespreading interval 608, with each element from this column being sent in one of N time slices 612 of the timespreading interval 608. Neglecting overhead effects, generally a complete N×N TFS matrix can be transmitted over N single time spreading intervals 622.
Attention is now directed to
Attention is now directed to
Thus every successive time slice, one element from the TFS matrix 2108 will be used to control the modulator 2104. In one embodiment of the OTFS method, the modulator 2104 includes modules 2132 and 2134 for separating the element into its real and imaginary components, modules 2142 and 2144 for chopping the resultant real and imaginary components, and filtering modules 2152 and 2154 for subsequently performing filtering operations. The filtered results are then used to control the operation of sin and cosine generators 2162 and 2164, the outputs of which are upconverted using a RF carrier in order to produce an analog radio waveform 2120. This waveform then travels to the receiver where it is demodulated and deconvolved as will be described below with reference to
In an alternative embodiment, diagonals of the TFS data matrix may be transmitted over a series of single timespreading intervals, one diagonal per single timespreading interval, so that N diagonals of the final N×N transmission matrix are transmitted over N time intervals. In other embodiments the order in which individual elements of the TFS transmission matrix [[U_{1}][D]][U_{2}] are transmitted across the communications link is determined by a transmit matrix or transmit vector.
In some embodiments, there may be some overhead to this basic model. Thus, for example, with some time padding (additional time slices or additional time spreading intervals), checksums or other verification/handshaking data, which could be transmitted in a nonconvolved manner, could be sent back by the receiver to the transmitter on a per timespreading interval, per N time spreading intervals, or even on a per time slice interval in order to request retransmission of certain parts of the TFS data matrix as needed.
Attention is now directed to
As previously discussed, in some embodiments, the first N×N time spreading matrix [U_{1}] may be constructed out of N rows of a cyclically shifted Legendre symbols or pseudorandom number of length N. That is, the entire N×N spreading matrix is filled with all of the various cyclic permutations of the same Legendre symbols. In some embodiments, this version of the [U_{1}] matrix can be used for spectrum spreading purposes and may, for example, instruct the transmitter to rapidly modulate the elements of any matrix that it affects rapidly over time, that is, with a chip rate that is much faster than the information signal bit rate of the elements of the matrix that the Legendre symbols are operating on.
In some embodiments, the second N×N spectral shaping matrix [U_{2}] can be a Discrete Fourier Transform (DFT) or an Inverse Discrete Fourier Transform (IDFT) matrix. These DFT and IDFT matrices can instruct the transmitter to spectrally shift the elements of any matrix that the DFT matrix coefficients act upon. Although many different modulation schemes may be used, in some embodiments, this modulation may be chosen to be an Orthogonal FrequencyDivision Multiplexing (OFDM) type modulation, in which case a modulation scheme such as quadrature amplitude modulation or phaseshift keying may be used, and this in turn may optionally be divided over many closelyspaced orthogonal subcarriers.
Often the actual choice of which coefficients to use for the first N×N timefrequency shifting matrix [U_{1}] and what coefficients to use for the second N×N spectral shaping matrix [U_{2}] may depend on the conditions present in the communications channel 320. If, for example, a communications channel 320 is subjected to a particular type of impairment, such as wide band noise, narrowband interference, impulse noise, Doppler shift, crosstalk and so on, then some first N×N timefrequency shifting matrices and some second N×N spectral shaping matrices will be better able to cope with these impairments. In some embodiments of the OTFS method, the transmitter and receiver will attempt to measure these channel impairments, and may suggest alternate types of first N×N timefrequency shifting matrices [U_{1}] and second N×N spectral shaping matrices to each [U_{2}] in order to minimize the data loss caused by such impairments.
Various modifications of the abovedescribed data transmission process represented by the matrix multiplication [[U_{1}][D]][U_{2}] are also within the scope of the present disclosure and are described below with reference to
Attention is now directed to
Attention is now directed to
 1: Receive ([P([U_{1}][D])][U_{2}])′ (stage 362)
 2: Perform a first left multiplication by the Hermitian matrix of the [U2] matrix [U_{2}^{H}], if one was used for transmission, thus creating P([U_{1}][D]) (stage 364).
 3: Inverse permute this replica by (P([U_{1}][D])P^{−1}, if a permutation was used during transmission, thus creating [U_{1}][D] (stage 368)
 4. Perform a second right multiplication by the Hermitian matrix of the [U_{1}] matrix [U_{1}^{H}], thus recreating [D] (stage 370).
As a consequence of noise and other impairments in the channel, use of information matrices and other noise reduction methods may be employed to compensate for data loss or distortion due to various impairments in the communications link. Indeed, it may be readily appreciated that one advantage of spreading out the original elements of the data frame [D] over a large range of times, frequencies, and spectral shapes as contemplated by embodiments of the OTFS method is that it becomes straightforward to compensate for the loss during transmission of information associated with a few of the many transmission times, frequencies and spectral shapes.
Although various deconvolution methods may used in embodiments of the OTFS method, the use of Hermitian matrices may be particularly suitable since, in general, for any Hermitian matrix [U^{H}] of a unitary matrix [U], the following relationship applies:
[U][U^{H}]=[I] where [I] is the identity matrix.
Communications links are not, of course, capable of transmitting data at an infinite rate. Accordingly, in one embodiment of the OTFS method the first N×N timefrequency shifting matrix ([U_{1}], the second N×N spectral shaping matrix ([U_{2}] (when one is used), and the elements of the data frame, as well as the constraints of the communications link (e.g. available bandwidth, power, amount of time, etc.), are chosen so that in balance (and neglecting overhead), at least N elements of the N×N TFS data matrix can be transmitted over the communications link in one timespreading interval. More specifically (and again neglecting overhead), one element of the N×N TFS data matrix will generally be transmitted during each time slice of each timespreading interval.
Given this rate of communicating data, then typically the entire TFS data matrix may be communicated over N timespreading intervals, and this assumption will generally be used for this discussion. However it should be evident that given other balancing considerations between the first N×N timefrequency shifting matrix, the second N×N spectral shaping matrix, and the elements of the data frame, as well as the constraints of the communications link, the entire TFS data matrix may be communicated in less than N timespreading intervals, or greater than N time spreading intervals as well.
As discussed above, the contents of the TFS data matrix may be transmitted by selecting different elements from the TFS data matrix, and sending them over the communications link, on a one element per time slice basis, over multiple spreading time intervals. Although in principle this process of selecting different elements of the TFS data matrix can be accomplished by a variety of different methods, such as sending successive rows of the TFS data matrix each single time spreading interval, sending successive columns of the TFS data matrix each successive time spreading interval, sending successive diagonals of the TFS data matrix each successive time spreading intervals, and so on, from the standpoint of communications link capacity, minimizing interference, and reducing ambiguity, some schemes are often better than others. Thus, often the [U_{1}] and [U_{2}] matrices, as well as the permutation scheme P, may be chosen to optimize transmission efficiency in response to various impairments in the communications link.
As shown in
 1: For each single timespreading interval, selecting N different elements of the TFS data matrix (often successive columns of the TFS matrix will be chosen)(stage 482).
 2: Over different time slices in the given time spreading interval, selecting one element (a different element each time slice) from the N different elements of the TFS data matrix, modulating this element, and transmitting this element so that each different element occupies its own time slice (stage 484).
 3: Receiving these N different replica elements of the transmitted TFS data matrix over different said time slices in the given time spreading interval (stage 486).
 4: Demodulating these N different elements of the TFS data matrix (stage 488).
 5. Repeating stages 482, 484, 486 and 488 up to a total of N times in order to reassemble a replica of the TFS data matrix at the receiver (stage 490).
This method assumes knowledge by the receiver of the first N×N spreading code matrix [U_{1}], the second N×N spectral shaping matrix [U_{2}], the permutation scheme P, as well as the particular scheme used to select elements from the TFS matrix to transmit over various periods of time. In one embodiment the receiver takes the accumulated TFS data matrix and solves for the original N×N data frame using standard linear algebra methods. It may be appreciated that because each original data symbol from the original data frame [D] has essentially been distributed over the entire TFS data matrix, it may not be possible to reconstruct an arbitrary element or symbol from the data [D] until the complete TFS data matrix is received by the receiver.
Attention is now directed to
As shown in
In order to decode or deconvolve the TFS matrix accumulated during stage 724, the digital OTFS data receiver 780 left multiplies, during a stage 726, the TFS matrix by the Hermitian matrix of the [U_{2}] matrix, i.e., [U_{2}^{H}], established at a stage 704. Next, the digital OTFS data receiver 780 performs, at a stage 728, an inverse permutation (P^{−1}) of the result of this left multiplication. The digital OTFS data receiver 780 then deconvolves the TFS matrix in order to reconstruct a replica 732 of the original data matrix [D] by, in a stage 730, right multiplying the result of stage 728 by the Hermitian of the original N×N timefrequency shifting matrix [U_{1}], i.e., [U_{1}^{H}], established at a stage 702. Because the reconstructed signal will generally have some noise and distortion due to various communications link impairments, various standard noise reduction and statistical averaging techniques, such as information matrices, may be used to assist in the reconstruction process (not shown). Each replicated frame 732 of each original data matrix [D] may be stored within digital data storage 782 (stage 740).
Attention is now directed to
Attention is now directed to
Attention is now directed to
Various modifications of the abovedescribed data reconstruction process are also within the scope of the present disclosure and are described below with reference to
Referring now to
Attention is now directed to
Consistent with one embodiment of the second form of the OTFS method, each element in the input frame of data [D] is multiplied by its corresponding unique waveform, thereby producing a series of N^{2 }weighted unique waveforms. Over one spreading time interval (generally composed of N time slices), all N^{2 }weighted unique waveforms corresponding to each data element in the frame of data [D] are simultaneously combined and transmitted. Further, in this embodiment a different unique basic waveform of length (or duration) of N time slices is used for each consecutive timespreading interval. The set of N unique basic waveforms (i.e., one for each of the N timespreading intervals) form an orthonormal basis. As may be appreciated, embodiments of the second form of the OTFS element contemplate that at least a part of [D] is transmitted within each of the N timespreading intervals.
To receive waveforms modulated and transmitted in accordance with this second form of the OTFS method, the received signal is (over each spreading interval of N time slices), correlated with the set of all N^{2 }waveforms previously assigned to each data element during the transmission process for that specific time spreading interval. Upon performing this correlation, the receiver will produce a unique correlation score for each one of the N^{2 }data elements (the receiver will have or be provided with knowledge of the set of N^{2 }waveforms respectively assigned by the transmitter to the corresponding set of N^{2 }data elements). This process will generally be repeated over all N timespreading intervals. The original data matrix [D] can thus be reconstructed by the receiver by, for each data element, summing the correlation scores over N timespreading intervals. This summation of correlation scores will typically yield the N^{2 }data elements of the frame of data [D].
Turning now to
In
In this alternative scheme or embodiment, the OTFS method can be understood as being a method of transmitting at least one frame of data ([D]) over a communications link, comprising: creating a plurality of timespectrumtone or spectralshape spreading codes operating over a plurality of timespreading intervals, each single timespreading interval being composed of at least one clock intervals; each timespectrumtone or spectralshape spreading code comprising a function of a first timefrequency shifting, a second spectral shaping, and a time spreading code or scanning and transmission scheme.
In an exemplary embodiment, OTFS modulation techniques may be employed to enable data to be sent from multiple users using multiple transmitters (here usually referred to as the multiple transmitter case) to be received by a single receiver. For example, assume multiple users “a”, “b”, “c”, and “d”, each desire to send a frame of data including N elements. Consistent with an embodiment of a multiuser OTFS transmission scheme, a conceptual N×N OTFS transmission matrix shared by multiple users may be created in the manner described below. Specifically, each given user packs their N elements of data into one column of an N×N data frame associated with such user but leaves the other columns empty (coefficients set to zero). The N×N data frame [D_{a}] associated with, and transmitted by, a user “a” may thus be represented as:
Similarly, the N×N data frame [D_{b}] associated with, and transmitted by, a user “b” may thus be represented as
And user “n” sends and N×N data frame [D_{n}]
Thus, transmission of the data frames [D_{a}], [D_{b}] . . . [D_{n}] respectively by the users “a”, “b” . . . “n” results in transmission of the conceptual N×N OTFS transmission matrix, with each of the users being associated with one of the columns of such conceptual transmission matrix. In this way each independent user “a”, “b” . . . “n” transmits its N data elements during its designated slot (i.e., column) within the conceptual N×N OTFS transmission matrix, and otherwise does not transmit information. This enables signals corresponding to the data frames [D_{a}], [D_{b}] . . . [D_{n}] to be received at the receiver as if the conceptual N×N OTFS transmission matrix was representative of a complete data frame sent by only a single transmitter. Once so received at the receiver, the received data frames [D_{a}], [D_{b}] . . . [D_{n}] effectively replicate the conceptual N×N OTFS transmission matrix, which may then be deconvolved in the manner discussed above.
The size of the tiles in
Multiple users that are using different transmitters (or simply multiple transmitters) may communicate over the same communications link using the same protocol. Here, each user or transmitter may, for example, select only a small number of data elements in the N^{2 }sized frame of data to send or receive their respective data. As one example, a user may simply select one column of the frame of data for their purposes, and set the other columns at zero. The user'"'"'s device will then compute TFS data matrices and send and receive them as usual.
As previously discussed, one advantage of the OTFS approach is increased resistance to Doppler shifts and frequency shifts. For example, in many cases the greater degree of time, frequency, and spectral shaping contemplated by the OTFS approach will largely mitigate any negative effects of such shifts due to the superior ability of OTFSequipped devices to function over an impaired communications link. In other cases, because the local impaired device can be identified with greater accuracy, a base station or other transmitting device can either send corrective signals to the impaired device, or alternatively shut off the impaired device.
As previously discussed, one advantage of the OTFS method is increased resistance to communications channel impairments. This resistance to impairments can be improved by further selecting the first N×N timefrequency shifting matrix and the second N×N spectral shaping matrix to minimize the impact of an aberrant transmitter; specifically, a transmitter suffering from Doppler shift or frequency shift on the elements of the TFS data matrix that are adjacent to the elements of the TFS data matrix occupied by the aberrant transmitter. Alternatively, the receiver may analyze the problem, determine if an alternate set of first N×N timefrequency shifting matrices and/or said second N×N spectral shaping matrices would reduce the problem, and suggest or command that corresponding changes be made to corresponding transmitter(s).
The OTFS method also enables more sophisticated tradeoffs to be made between transmission distance, transmitter power, and information data rate than is possible to be made using conventional modulation techniques. This increased flexibility arises in part because each symbol is generally spread over a larger number of intervals relative to the case in which conventional techniques are employed. For example, in conventional timedivision multiplexed communication systems the power per symbol transmitted must be quite high because the symbol is being transmitted over only one time interval. In conventional spread spectrum communication systems, the symbol is being transmitted over essentially N intervals, and the power per interval is correspondingly less. Because the OTFS method transmits a bit or symbol of information over N^{2 }different modalities (e.g. waveforms, times), the power per modality is much less. Among other things, this means that the effect of impulse noise, that would in general only impact a specific waveform over a specific time interval, will be less. It also means that due to increased number of signal transmission modalities (waveforms, times) enabled by the OTFS method, there are more degrees of freedom available to optimize the signal to best correspond to the particular communications link impairment situation at hand.
Attention is now directed to
Since a portion of the transmitted signal 2702 is a cyclically time shifted waveform, a time deconvolution device 2714 at the receiver, such as the post—equalizer 480 of
Thus, the direct signals 2804 impinging upon the receiver 2806 will, in this example, not be frequency shifted. However the Dopplershifted wireless signals 2808 that bounce off of the wireless reflector, here again building 2807, will echo off in a higher frequency shifted form. These higher frequency shifted “echo” reflections 2810 also still have to travel a longer distance to reach receiver 2806, and thus also end up being time delayed as well. As a result, receiver 2806 receives a signal 2812 that is distorted due to the summation of the direct signal 2804 with the time and frequency shifted echo waveforms 2810.
However, as was described above, the OTFS techniques described herein may utilize the transmission of cyclically time shifted and frequency shifted waveforms. Accordingly, a time and frequency deconvolution device 2814 (alternatively a time and frequency adaptive equalizer such as the OTFS demodulator 460 and the OTFS postequalizer 480 of
The net effect of both time and frequency deconvolutions, when applied to transmitters, receivers, and echo sources that potentially exist at different distances and velocities relative to each other, is to allow the receiver to properly interpret the impaired signal. Here, even if the energy received in the primary signal is too low to permit proper interpretation, the energy from the time and/or frequency shifted versions of the signals can be added to the primary signal upon the application of appropriate time and frequency offsets or deconvolution parameters to signal versions, thereby resulting in a less noisy and more reliable signal at the receiver. Additionally, the time and frequency deconvolution parameters can contain useful information as to the relative positions and velocities of the echo location(s) relative to the transmitter and receiver, as well as the various velocities between the transmitter and receiver, and can also help the system characterize some of the signal impairments that occur between the transmitter and receiver.
Thus, in some embodiments the OTFS systems described herein may also provide a method to provide an improved receiver where, due to either one or a combination of echo reflections and frequency offsets, multiple signals associated with such reflections and offsets result in the receiver receiving a time and/or frequency convolved composite signal representative of time and/or frequency shifted versions of the N^{2 }summationsymbolweighted cyclically time shifted and frequency shifted waveforms. Here, the improved receiver will further time and/or frequency deconvolve the time and/or frequency convolved signal to correct for such echo reflections and the resulting time and/or frequency offsets. This will result in both time and frequency deconvolved results (i.e. signals, typically of much higher quality and lower signal to noise ratio), as well as various time and frequency deconvolution parameters that, as will be discussed, are useful for a number of other purposes.
Before going into a more detailed discussion of other applications, however, it is useful to first discuss the various waveforms in more detail.
Embodiments of the OTFS systems and methods described herein generally utilize waveforms produced by distributing a plurality of data symbols into one or more N×N symbol matrices, and using these one or more N×N symbol matrices to control the signal modulation of a transmitter. Here, for each N×N symbol matrix, the transmitter may use each data symbol to weight N waveforms, selected from an N^{2}sized set of all permutations of N cyclically time shifted and N cyclically frequency shifted waveforms determined according to an encoding matrix U, thus producing N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms for each data symbol. This encoding matrix U is chosen to be an N×N unitary matrix that has a corresponding inverse decoding matrix U^{H}. The method will further, for each data symbol in the N×N symbol matrix, sum the N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms, producing N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms. The transmitter will transmit these N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms, structured as N composite waveforms, over any combination of N time blocks or frequency blocks.
As discussed above, various waveforms can be used to transmit and receive at least one frame of data [D] (composed of a matrix of up to N^{2 }data symbols or elements) over a communications link. Here each data symbol may be assigned a unique waveform (designated a corresponding waveform) derived from a basic waveform.
For example, the data symbols of the data matrix [D] may be spread over a range of cyclically varying time and frequency shifts by assigning each data symbol to a unique waveform (corresponding waveform) which is derived from a basic waveform of length N time slices (in embodiments described herein the set of N time slices correspond to the time required to transmit this waveform, also referred to as a time block), with a data symbol specific combination of a time and frequency cyclic shift of this basic waveform.
In one embodiment each symbol in the frame of data [D] is multiplied by its corresponding waveform, producing a series of N^{2 }weighted unique waveforms. Over one spreading time interval (or time block interval), all N^{2 }weighted unique waveforms corresponding to each data symbol in the frame of data [D] are simultaneously combined and transmitted. Further, a different unique basic waveform of length (or duration) of one time block (N time slices) may be used for each consecutive timespreading interval (consecutive time block). Thus a different unique basic waveform corresponding to one time block may be used for each consecutive timespreading interval, and this set of N unique waveforms generally forms an orthonormal basis. Essentially, each symbol of [D] is transmitted (in part) again and again either over all N time blocks, or alternatively over some combination of time blocks and frequency blocks (e.g. assigned frequency ranges).
To receive data over each block of time, the received signal is correlated with a corresponding set of all N^{2 }waveforms previously assigned to each data symbol by the transmitter for that specific time block. Upon performing this correlation, the receiver may produce a unique correlation score for each one of the N^{2 }data symbols. This process will be repeated over some combination of time blocks and frequency blocks until all N blocks are received. The original data matrix [D] can thus be reconstructed by the receiver by summing, for each data symbol, the correlation scores over N time blocks or frequency blocks, and this summation of the correlation scores will reproduce the N^{2 }data symbols of the frame of data [D].
Note that in some embodiments, some of these N time blocks may be transmitted nonconsecutively, or alternatively some of these N time blocks may be frequency shifted to an entirely different frequency range, and transmitted in parallel with other time blocks from the original set of N time blocks in order to speed up transmission time. This is discussed later and in more detail in reference to
In order to enable focus to be directed to the underlying cyclically time shifted and cyclically shifted waveforms, detailed aspects of one embodiment of the OTFS methods described above may be somewhat generalized and also discussed in simplified form. For example, the operation of selecting from an N^{2 }set of all permutations of N cyclically time shifted and N cyclically frequency shifted waveforms may correspond, at least in part, to an optional permutation operation P as well as to the other steps discussed above. Additionally, the N^{2 }set of all permutations of N cyclically time shifted and N cyclically frequency shifted waveforms may be understood, for example, to be at least partially described by a Discrete Fourier transform (DFT) matrix or an Inverse Discrete Fourier Transform matrix (IDFT). This DFT and IDFT matrix can be used by the transmitter, for example, to take a sequence of real or complex numbers and modulate them into a series of different waveforms.
Considering now a particular example, individual rows of a DFT matrix (e.g., the DFT matrix of
where, for an N×N DFT matrix [X}, X_{j}^{k }is the coefficient of the Fourier vector in row k column j of the DFT matrix, and N is the number of columns. The products of this Fourier vector may be considered to represent one example of a manner in which the various time shifted and frequency shifted waveforms suitable for use in the OTFS system may be generated.
For example and as mentioned previously,
Although previously discussed,
As described above with reference to
Referring now to
As yet another alternative, some of the various waveform time blocks may be frequency transposed to entirely different frequency bands or ranges 2912, 2914, 2916. This can speed up transmission time, because now multiple waveform time blocks can now be transmitted at the same time as different frequency blocks. As shown in time/frequency offset tiles 2918 and 2920, such multiple frequency band transmissions can also be done on a contiguous, sparse contiguous, contiguous interleaved, or sparse contiguous interleaved manner. Here 2922 and 2928 represent one time block at a first frequency range 2912, and 2924 and 2930 represent the next time block at the frequency range 2912. Here the various frequency ranges 2912, 2914, and 2916 can be formed, as will be described shortly, by modulating the signal according to different frequency carrier waves. Thus, for example, frequency range or band 2912 might be transmitted by modulating a 1 GHz frequency carrier wave, frequency range or band 2914 might be transmitted by modulating a 1.3 GHz frequency carrier wave, and band 2915 might be transmitted by modulating a 1.6 GHz frequency carrier wave, and so on.
Stated differently, the N composite waveforms, themselves derived from the previously discussed N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms, may be transmitted over at least N time blocks. These N time blocks may be either transmitted consecutively in time (e.g. 2902, 2904) or alternatively transmitted timeinterleaved with the N time blocks from a second and different N×N symbol matrix.
For both
Attention is now again directed to
In some embodiments the transmitter 2100 may further implement a preequalization operation, typically performed by the preequalizer 410 of
In
In one embodiment the equalizer 3102 produces a set of equalization parameters 3108 during the process of equalizing the distorted waveform. For example, in the simple case where the original waveform was distorted by only a single echo reflection offset by time t_{offset}, and by the time the original waveform and the t_{offset }echo waveform reach the receiver, the resulting distorted signal 3100 may be, for example, about 90% original waveform and 10% t_{offset }echo waveform, then the equalization parameters 3108 can output both the 90% original and 10% echo signal mix, as well as the t_{offset }value. Typically, of course, the actual distorted signal 3100 could consist of a number of various time and frequency offset components, and here again, in addition to cleaning this distortion, the equalizer 3102 can also report the various time offsets, frequency offsets, and percentage mix of the various components of signal 3100 to the transmitter and/or the receiver.
As previously discussed in
In some embodiments, for example in network systems where there may be multiple transmitters and potentially also multiple receivers, it may be useful to transmit the data from the various transmitters using more than one encoding method. Here, for example, a first set of N time blocks may transmit data symbols originating from a first N×N symbol matrix from a first transmitter using a first unitary matrix [U_{1}]. A second set of N time blocks may transmit data symbols originating from a second N×N symbol matrix from a second transmitter using a second unitary matrix [U_{2}]. Depending on the embodiment, [U_{1}] and [U_{2}] may be identical or different. Because the signals originating from the first transmitter may encounter different impairments (e.g. different echo reflections, different frequency shifts), some schemes of cyclically time shifted and cyclically frequency shifted waveforms may operate better than others. Thus, these waveforms, as well as the unitary matrices [U_{1}] and [U_{2}], may be selected based on the characteristics of these particular echo reflections, frequency offsets, and other signal impairments of the system and environment of the first transmitter, the second transmitter and/or the receiver.
As an example, a receiver configured to implement equalization in accordance with
In some cases, before transmitting large amounts of data, or any time as desired, a given transmitter and receiver may choose to more directly test the various echo reflections, frequency shifts, and other impairments of the transmitter and receiver'"'"'s system and environment. This can be done, by, for example having the transmitter send a test signal where the plurality of data symbols are selected to be test symbols known to the receiver (e.g., the receiver may have stored a record of these particular test symbols). Since in this case the receiver will be aware of exactly what sort of signal it should receive in the absence of any impairment, the equalizer 3102 will generally be able to provide even more accurate time and frequency equalization parameters 3108 for use by the receiver relative to the case in which the receiver lacks such awareness. Thus, in this case the equalization parameters provide even more accurate information relating to the characteristics of the echo reflections, frequency offsets, and other signal impairments of the system and environment of the applicable transmitter(s) and the receiver. This more accurate information may be used by the receiver to suggest or command that the applicable transmitter(s) shift to use of communications schemes (e.g., to U matrices) more suitable to the present situation.
In some embodiments, when the transmitter is a wireless transmitter and the receiver is a wireless receiver, and the frequency offsets are caused by Doppler effects, the more accurate determination of the deconvolution parameters, i.e. the characteristics of the echo reflections and frequency offsets, can be used to determine the location and velocity of at least one object in the environment of the transmitter and receiver.
This section includes a description of a number of exemplary OTFS equalization techniques capable of being implemented consistent with the general OTFS equalization approach and apparatus discussed above. However, prior to describing such exemplary techniques, a summary of aspects of transmission and reception of OTFSmodulated signals is given in order to provide an appropriate context for discussion of these OTFS equalization techniques.
Turning now to such a summary of OTFS signal transmission and reception, consider the case in which a microprocessorcontrolled transmitter packages a series of different symbols “d” (e.g. d_{1}, d_{2}, d_{3 }. . . ) for transmission by repackaging or distributing the symbols into various elements of various N×N matrices [D]. In one implementation such distribution may, for example, include assigning d_{1 }to the first row and first column of the [D] matrix (e.g. d_{1}=d_{0,0}), d_{2 }to the first row, second column of the [D] matrix (e.g. d_{2}=d_{0,1}) and so on until all N×N symbols of the [D] matrix are full. Here, if the transmitter runs out of “d” symbols to transmit, the remaining [D] matrix elements can be set to be 0 or other value indicative of a null entry.
The various primary waveforms used as the primary basis for transmitting data, which here will be called “tones” to show that these waveforms have a characteristic sinusoid shape, can be described by an N×N Inverse Discrete Fourier Transform (IDFT) matrix [W], where for each element w in [W],
or alternatively W_{j,k}=e^{ijθ}^{k }or w_{j,k}=[e^{iθ}^{k}]^{j}. Thus the individual data elements d in [D] are transformed and distributed as a combination of various fundamental tones w by a matrix multiplication operation [W]*[D], producing a tone transformed and distributed form of the data matrix, here described by the N×N matrix [A], where [A]=[W]*[D].
To produce N cyclically time shifted and N cyclically frequency shifted waveforms, the tone transformed and distributed data matrix [A] is then itself further permuted by modular arithmetic or “clock” arithmetic, thereby creating an N×N matrix [B], including each element b of [B], b_{i,j}=a_{i,(i+j)mod N}. This can alternatively be expressed as [B]=Permute([A])=P(IDFT*[D]). Thus the clock arithmetic controls the pattern of cyclic time and frequency shifts.
The previously described unitary matrix [U] can then be used to operate on [B], producing an N×N transmit matrix [T], where [T]=[U]*[B], thus producing an N^{2 }sized set of all permutations of N cyclically time shifted and N cyclically frequency shifted waveforms determined according to an encoding matrix [U].
Put alternatively, the N×N transmit matrix [T]=[U]*P(IDFT*[D]).
Then, typically on a per column basis, each individual column of N is used to further modulate a frequency carrier wave (e.g. if transmitting in a range of frequencies around 1 GHz, the carrier wave will be set at 1 GHz). In this case each Nelement column of the N×N matrix [T] produces N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms for each data symbol. Effectively then, the transmitter is transmitting the sum of the N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms from one column of [T] at a time as, for example, a composite waveform over a time block of data. Alternatively, the transmitter could instead use a different frequency carrier wave for the different columns of [T], and thus for example transmit one column of [T] over one frequency carrier wave, and simultaneously transmit a different column of [T] over a different frequency carrier wave, thus transmitting more data at the same time, although of course using more bandwidth to do so. This alternative method of using different frequency carrier waves to transmit more than one column of [T] at the same time will be referred to as frequency blocks, where each frequency carrier wave is considered its own frequency block.
Thus, since the N×N matrix [T] has N columns, the transmitter will transmit the N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms, structured as N composite waveforms, over any combination of N time blocks or frequency blocks, as previously shown in
On the receiver side, the transmit process is essentially reversed. Here, for example, a microprocessor controlled receiver would of course receive the various columns [T] (e.g., receive the N composite waveforms, also known as the N symbolweighted cyclically time shifted and cyclically frequency shifted waveforms) over various time blocks or frequency blocks as desired for that particular application. In cases in which sufficient bandwidth is available and time is of the essence, the transmitter may transmit the data as multiple frequency blocks over multiple frequency carrier waves. On the other hand, if available bandwidth is more limited, and/or time (latency) is less critical, then the transmit will transmit and the receiver will receive over multiple time blocks instead.
During operation the receiver may effectively tune into the one or more frequency carrier waves, and over the number of time and frequency blocks set for the particular application, eventually receive the data or coefficients from the original N×N transmitted matrix [T] as an N×N receive matrix [R]. In the general case [R] will be similar to [T], but may not be identical due to the existence of various impairments between the transmitter and receiver.
The microprocessor controlled receiver then reverses the transmit process as a series of steps that mimic, in reverse, the original transmission process. The N×N receive matrix [R] is first decoded by inverse decoding matrix [U^{H}], producing an approximate version of the original permutation matrix [B], here called [B^{R}], where [B^{R}]=([U^{H}]*[R]).
The receiver then does an inverse clock operation to back out the data from the cyclically time shifted and cyclically frequency shifted waveforms (or tones) by doing an inverse modular mathematics or inverse clock arithmetic operation on the elements of the N×N [B^{R}] matrix, producing, for each element b^{R }of the N×N [B^{R}] matrix, a_{i,j}^{R}=b_{i,(j−i)mod N}^{R}. This produces a decyclically time shifted and decyclically frequency shifted version of the tone transformed and distributed form of the data matrix [A], which may hereinafter be referred to as [A^{R}]. Put alternatively, [A^{R}]=Inverse Permute ([B^{R}]), or [A^{R}]=P^{−1}([U^{H}]*[R]).
The receiver then further extracts at least an approximation of the original data symbols d from the [A^{R}] matrix by analyzing the [A] matrix using an N×N Discrete Fourier Transform matrix DFT of the original Inverse Fourier Transform matrix (IDFT).
Here, for each received symbol d^{R}, the d^{R }are elements of the N×N received data matrix [D^{R}] where [D^{R}]=DFT*A^{R}, or alternatively [D^{R}]=DFT*P^{−1}([U^{H}]*[R]).
Thus the original N^{2 }summationsymbolweighted cyclically time shifted and cyclically frequency shifted waveforms are subsequently received by a receiver which is controlled by the corresponding decoding matrix U^{H }(also represented as [U^{H}]) The processor of the receiver uses this decoding matrix [U^{H}] to reconstruct the various transmitted symbols “d” in the one or more originally transmitted N×N symbol matrices [D] (or at least an approximation of these transmitted symbols).
Turning now to a discussion of various exemplary OTFS equalization techniques, there exist at least several general approaches capable of being used correct for distortions caused by the signal impairment effects of echo reflections and frequency shifts. One approach leverages the fact that the cyclically time shifted and cyclically frequency shifted waveforms or “tones” form a predictable timefrequency pattern. In this scheme a deconvolution device situated at the receiver'"'"'s front end may be straightforwardly configured to recognize these patterns, as well as the echoreflected and frequency shifted versions of these patterns, and perform the appropriate deconvolutions by a pattern recognition process. Alternatively the distortions may be mathematically corrected using software routines, executed by the receiver'"'"'s processor, designed to essentially determine the various echo reflected and frequency shifting effects, and solve for these effects. As a third alternative, once, by either process, the receiver determines the time and frequency equalization parameters of the communication media'"'"'s particular time and frequency distortions, the receiver may transmit a command to the transmitter to instruct the transmitter to essentially precompensate or preencode, e.g., by using a preequalizer such as the preequalizer 410 of
An exemplary equalization mechanism associated with the OTFS modulation, which is inherently two dimensional, is discussed below. This is in contrast to its one dimensional counterpart in conventional modulation schemes such as, for example, OFDM and TDMA.
Assume that the input symbol stream provided to an OTFS transmitter is a digital function Xε(R_{d}×R_{d}) with values in specific finite constellation C⊂ (either QPSK or higher QAM'"'"'s, for example). This transmitter modulates this input stream into an analog signal φ^{Tx,Pass}, which is then transmitted. During the transmission φ^{Tx,Pass }undergo a multipath channel distortion. The distorted passband signal φ^{Tx,Pass }arrives at the OTFS receiver and is demodulated back into a digital function Yε(R_{d}×R_{d}), which may be referred to herein as the output stream. The locality properties of OTFS modulation imply that the net effect of the multipath channel distortion is given by a cyclic two dimensional convolution with a two dimensional channel impulse response. See
Referring to
In what follows an appropriate equalization mechanism will be described. To this end, it will be convenient to enumerate the elements of the digital time axis by 0, 1, . . . , N1 and to consider the input and output streams X and Y, respectfully, as sequences of functions:
X=(X(k)ε(R_{d}):k=0, . . . ,N−1),
Y=(Y(k)ε(R_{d}):k=0, . . . ,N−1),
where X(k)(i)=X(k,i) and Y(k)(i)=Y(k,i), for every k=0, . . . , N−1 and iεR_{d}.
Furthermore, for purposes of explanation it will be assumed that the time index k is infinite in both directions, that is, kε, the digital time direction is linear, and the digital frequency direction is cyclic. Under these conventions the relation between the output stream and the input stream can be expressed by the following Equation (1.1):
where:
 C=(C(l)ε(R_{d}): l=L_{C}, . . . , R_{C}) are the channel impulse taps. Typically, L_{C}68 ^{<0 }and R_{C}ε^{>0}. The number n_{C}=R_{C}−L_{C}+1 may be referred to herein as the memory length of the channel. The operation * in (4.1) stands for one dimensional cyclic convolution on the ring R_{d}.
 (k)ε(0, N_{0}−Id_{N}) is a complex Gaussian N dimensional vector with mean 0 and covariance matrix N_{0}·Id, representing the white Gaussian noise.
Referring now to
The adaptive decision feedback equalizer 3250 may, in some embodiments, operate according to the function:
An exemplary decision feedback LMS equalizer adapted to the relation expressed in Equation (1.1) will now be described under the condition that the carrier frequency is locked between the transmitter and receiver, that is, W_{Tx}=W_{Rx}. An adaptation of the equalizer under the condition of the existence of a nonzero discrepancy, i.e., ΔW≠0, will subsequently be described. In one aspect, the equalizer incorporates a forward filter and a feedback filter as follows:
F=(F(l)ε(R_{d}):l=L_{F}, . . . ,R_{F}), Forward filter:
B=(B(l)ε(R_{d}):l=L_{B}, . . . ,−1), Feedback filter:
where, typically L_{F}, L_{B}ε^{<0 }and R_{F}ε^{0 }satisfy L_{F}, L_{B}≧L_{C }and R_{F}R_{C}. In fact, both filters depend on additional parameter kε designating the present point on the digital time axis, hence, the complete notation for the filter taps is F_{k }(l) and B_{k }(l). However, for the sake of presentation this additional index will generally be omitted and will only be included when necessary. A soft estimator is defined as follows:
where X^{h}(k+l) is the past hard estimation of the past data vector X(k+l), l=L_{B}, . . . , −1, defined as the quantization X^{h}(k)=Q(X^{8}(k)), that is:
In one aspect, a closed formula may be used to determine the forward and feedback filter taps of the decision feedback equalizer expressed in terms of the channel impulse response. In this case the forward filter taps are computed without regard to feedback and then the feedback filter taps are determined.
First fix k=0 and let X^{8 }denote the following soft estimator for the vector X (0), which depends only upon the forward filter taps:
In what follows it is assumed that X(k)˜C(0, P·Id_{N}), for every kε. Later this condition may be replaced by the condition that X(k)˜C(0, P·Id_{N}), for k≧0 and X(k)=0 for k<0, which is more adapted to the choice of X^{8}. We denote by Err=Err (0) the soft error term:
Err=X^{8}−X(0). 1.8
We consider the cost function:
U(F)=∥Err∥^{2}=∥X^{8}−X(0)∥^{2 }
where the expectation is taken over the probability distribution of the input stream X and the additive white Gaussian noise . The optimal filter F^{opt }is defined as:
therefore it satisfies the following system of linear equations:
∇_{F(l)}U(F^{opt})=0,l=L_{F}, . . . ,R_{F}. 1.9
The formula for the gradient ∇_{F(l)}U is an averaged version of (1.6), that is:
We first compute the term [X (0)*Y(l)★] and then the term [X^{6}*Y(l)★]. Developing the expression X(0)*Y(l))★ we obtain:
We observe that [X(0)*X(k)★]=0 when l≠0 and [X(0)*X(0)★]=NP•δ_{w=0}, hence we conclude that:
[X(0)*Y(l)^{★}]=NP·δ_{w=0}*C(l)^{★}=NP·C(l)^{★}. 1.11
Next, we compute the term [X^{δ}*Y(l)★]:
Developing the expression for Y(l′)*Y(l) we obtain:
Denote
Taking the expectation of both sides of (1.13) we obtain the following explicit formula for R(l,l′):
where in the computation of R(l,l′) we use the following conditions on the mean of the specific terms in (1.13):
[X(k)*X(k′)]=NP·δ_{k=k′},
[(k)*(k′)]=NN_{0}·δ_{k=k′},
[Additional Terms]=0.
Combining (1.9), (1.10), (1.11) and (1.12) we conclude that the optimal filter F^{opt }satisfies the following system of linear equations:
Finally, system (1.15) can be reduced to N systems of n_{F}=R_{F}−L_{F}=1 scalar equations as follows. Applying a DFT to both sides of (1.15) we obtain:
where stands for the DFT of the corresponding function and • stands for pointwise multiplication of functions in (R_{d}) as we recall that DFT interchanges convolution with pointwise multiplication and ★ with complex conjugation. Now observe that each function valued equation in (1.16) decouples into n_{F }scalar valued equations by evaluating both sides on each element of the ring R_{d}. Explicitly, if we number the elements in R_{d }by 0, 1, 2, . . . , N−1 we end up with the following scalar valued system of equations:
for every i=0, N−1. In more concrete matrix form (1.17) looks like:
for every i=0, N−1. We conclude the discussion by considering the case when the input stream satisfies X(k)=0 for k<0 which is the case adapted to (feedback) subtraction of the past interference. In this scenario, the optimal forward filter F^{opt }satisfies a system of the form (1.15) with the “matrix coefficients” R(l,l′) taking the form:
The optimal feedback filter taps B^{opt}(l), l=L_{B}, . . . , −1 can be computed from the forward and channel taps according to the following formula:
The justification of Formula (1.20) proceeds as follows. Fix an input vector X(l_{0}) for some specific l_{0}=L_{B}, . . . , −1. Subtracting its interference C(l′)*X(l_{0}) from each term Y(l_{0}+l′), we obtain an “interference free” sequence {tilde over (Y)}(l), l=L_{F}, . . . , R_{F}. Now, applying the forward filter F^{opt }to the sequence {tilde over (Y)}(l) we obtain an estimator for X(0) given by
which concludes the justification.
In an alternative aspect, a closed formula of the optimal forward and feedback filter taps of the decision feedback equalizer may be expressed in terms of the channel impulse response. In this regard we conduct the computation in the stochastic setting where we assume that X (k)˜C(0, P•Id_{N}), for every kε. We denote by X^{8 }the following soft estimator for the vector X (0):
We denote by Err=Err (0) the soft error term:
Err=X^{8}−X(0). 1.22
We consider the cost function:
U(F,B)=∥Err∥^{2}=X^{8}−X(0)^{2},
where the expectation is taken over the probability distribution of the input stream X and the additive white Gaussian noise . The optimal filters F^{opt}, B^{opt }are defined as:
therefore they satisfy the following system of linear equations:
∇_{F(l)}U(F_{opt},B^{opt})=0,l=L_{F}, . . . ,R_{F}. 1.23
∇_{B(l)}U(F^{opt},B^{opt})=0,l=L_{B}, . . . ,−1,
where the gradients are given by:
∇_{F(l)}U=[Err*Y(l)^{★}],l=L_{F}, . . . ,R_{F}, 1.24
∇_{B(l)}U=−[Err*X(l)*],l=L_{B}, . . . ,−1.
First we write explicitly the first system ∇_{F(l)}U (F^{opt}, B^{opt})=0. Expanding the term [Err*Y(l)★], we obtain:
Direct computation reveals that:
[X(0)*Y(l)^{★}]=NP·C(l)^{★},
[X(l′)*Y(l)^{★}]=NP·C(l−l′),
[Y(l′)*Y(Y(l)^{★}]=NP·R_{1}(l,l′),
where:
Thus, the first system of equations amounts to:
Next we write explicitly the system ∇_{B(l)}U(F^{opt}, B^{opt})=0. Expanding the term [Err*X(l)★], we obtain:
Direct computation reveals that:
[Y(l′)*X(l)^{★}]=NP·C(l′−l),
[X(l′)*X(l)^{★}]=NP·δ_{l=l′}·δ_{w=0},
[X(0)*X(l)^{★]=}0,
Thus, the second system of equations amounts to:
Using Equation (1.26), the optimal feedback filter taps may be expressed in terms of the channel taps and the optimal forward filter taps as:
Substituting the right hand side of (1.27) in (1.25) enables the optimal forward filter taps to be determined by finding the solution of the following linear system:
where:
As a final note, we denote R(l,l′)=R_{1}(l,l′)−R_{2}(l,l′) and write the system (1.28) in the following form:
System (1.29) can be reduced to N systems of n_{F}=R_{F}−L_{F}=1 scalar equations as follows. Applying a DFT to both sides of (1.15) we obtain:
Where stands for the DFT of the corresponding function and • stands for pointwise multiplication of functions in (R_{d}), since the DFT interchanges convolution with pointwise multiplication and ★ with complex conjugation. It is now observes that each function valued equation in (1.30) decouples into n_{F }scalar valued equations by evaluating both sides on each element of the ring R_{d}. Explicitly, numbered the elements in R_{d }as 0, 1, 2, . . . , N−1 results in the following scalar valued system of equations:
for every i=0, . . . , N−1. In more concrete matrix form (1.31) looks like:
for every i=0, . . . , N−1.
An exemplary channel acquisition component of the OTFS modulation scheme will now be described. To this end, we number the elements of R_{d }by 0, 1, . . . , N−1. For the channel acquisition, a rectangular strip [0, R_{C}−2L_{C}]×[0, N] is devoted in the time frequency plane. The value of the input stream X at this strip is specified to be:
The complement of this stream will generally be devoted to data.
As mentioned previously, the forward and feedback taps of the decision feedback equalizer depend on the index k and change slowly as k varies. We proceed to describe herein an exemplary tracking mechanism based on gradient correction with respect to an appropriate quadratic cost function. We denote by Err (k) the soft error term at the k step:
Err(k)=X^{8}(k)−X^{h}(k)ε(R_{d}), 1.4
where theoretically this error should be taken with respect to the true data vector X (k) (true decisions); however, in an exemplary embodiment the error is taken with respect to the hard estimator X^{h }(k) (hard decisions) as specified in equation (1.4). We define the following cost function U, taking as arguments the forward and feedback filter taps:
U(F,B)=∥Err(k)∥^{2}=∥X^{8}(k)−X^{h}(k)∥^{2}, 1.5
where ∥−∥ is the norm associated with the standard Hermitian inner product (−,−) on (R_{d}). Note that, in fact, the cost function depends on the index k, however, for the sake of brevity we omit this index from the notation. Next, we compute the gradients ∇_{F(l)}U, l=L_{F}, . . . , B_{F }and ∇_{B(l)}U, l=L_{B}, −1 with respect to the Euclidean inner product 2 Re (−,−) on (R_{d}). (considered as a real vector space). The formulas for the gradients are:
∇_{F(l)}=∇_{F(l)}U=Err(k)*Y(k+l)^{★},l=L_{F}, . . . ,R_{F}, 1.6
∇_{B(l)}=∇_{B(l)}U=−Err(k)*X^{h}(k+l)^{★},l=L_{B}, . . . ,−1,
where ★ stands for the star operation on the convolution algebra (R_{d}), given by f★ (i)=
The correction of the taps at the k step is obtained by adding a small increment in the (inverse) gradient direction, that is:
F_{k+1}(l)=F_{k}(l)−μ·∇_{F(l)},l=L_{F}, . . . ,R_{F},
B_{k+1}(l)=B_{k}(l)−μ·∇_{F(l)},l=L_{B}, . . . ,−1
for an appropriately chosen positive real number μ<<1. The optimal value μ_{opt }of the small parameter μ is given by:
A formal development of the quadratic expression U(F+μ·∇_{F}, B+μ·∇_{B}) in the parameter μ reveals that:
where Hess (∇_{F}, ∇_{B}) stands for:
and (F,F) and (B,B) stand for:
If we denote b=2 Re (F,F)+2 Re (B,B) and a=Hess (∇_{F}, ∇_{B}) then the standard formula for the minimum of a parabola imply that μ_{opt }is given by:
The quantizer portion of the adaptive decision feedback equalizer 3214 then acts to “round” the resulting signal to the nearest quantized value so that, for example, the symbol “1” after transmission, once more appears on the receiving end as “1” rather than “0.999”.
As previously discussed, an alternative mathematical discussion of the equalization method, particularly suitable for step 802B, is described in provisional application 61/615,884, the contents of which are incorporated herein by reference.
Attention is now directed to
In the embodiment of
An RF down converter 3725 demodulates the second received signal and passes the demodulated received second data stream D_{r }to a first OTFS decoder 37301 and a second OTFS decoder 37302. The first OTFS decoder 37302 decodes the received second signal using the base t matrix that was used to transmit the first data stream. The second OTFS decoder 37302 decodes the echocanceled data stream using a base r matrix that the other transmitter used to encode the second data stream. The output of the first OTFS decoder 37301 is fed back as a residual error signal to the echo canceller 3705 in order to tune the two dimensional estimate of the reflected echoes channel. The output of the second OTFS decoder 37302 is an estimate of the second data stream from the other transmitter. The capability to obtain an estimate of the echo channel in both frequency and time is a significant advantage of the OTFS technique, and facilitates fullduplex communication over a common frequency band in a manner not believed to be possible using prior art methods.
The OTFS receiver 3800 includes a pair of feedforward and feedback equalizers comprising first and second feed forward equalizers 38201 and 38202, first and second feedback equalizers 38351 and 38352, and first and second slicers 38251 and 38252. First and second subtractors 38301 and 38302 calculate first and second residual error signals 38401 and 38402 that are used by respective ones of the feed forward equalizers 3820 and the feedback equalizers 3835 in order to optimize two dimensional time/frequency shift channel models.
A pair of cross talk cancellers 38451 and 38452 also use the residual error signals 38401 and 38402, respectively, in order to optimize estimates of the first received data signal and the second received data signal in order to subtract each signal at subtractors 38151 and 38152. In this way, the cross talk from one data signal to the other is minimized. As with the full duplex OTFS transceiver 3700 of
Attention is now directed to
In the embodiment of
Turning now to
Referring to
The timefrequencyspace decision feedback equalizer 4000 advantageously enables the separation of signals within an OTFS communication system in a manner that substantially maximizes utilization of the available bandwidth. Such signal separation is useful in several contexts within an OTFS communication system. These include separation, at a receiver fed by multiple colocated or noncolocated antennas, of signals transmitted by a set of colocated or noncolocated antennas of a transmitter. In addition, the timefrequencyspace decision feedback equalizer 4000 enables the separation, from signal energy received from a remote transmitter, of echoes received by a receive antenna in response to transmissions from a nearby transmit antenna. This echo cancellation may occur even when the transmit and receive signal energy is within the same frequency band, since the twodimensional channelmodeling techniques described herein enable accurate and stationary representations of the both the echo channel and the channel associated with the remote transmitter. Moreover, as is discussed below the signal separation capability of the disclosed timefrequencyspace decision feedback equalizer enables deployment of OTFS transceivers in a mesh configuration in which neighboring OTFS transceivers may engage in full duplex communication in the same frequency band with other such transceivers in a manner transparent to one another.
Again with reference to
At each of the M antenna instances associated with an OTFS receiver, each entry within the twodimensional array of received signal energy being collected will typically include a contribution from each of the N transmit antenna instances involved in transmitting such signal energy. That is, each of the M receive antenna instances collects a mixture of the twodimensional, timefrequency planes of information separately sent by each of the N transmit antenna instances. Thus, the problem to be solved by the equalizer 4000 may be somewhat simplistically characterized as inversion of the N×M “coupling matrix” representative of the various communication channels between the N OTFS transmit antenna instances and the M OTFS receive antenna instances.
In one embodiment each of the N transmit antenna instances sends a pilot signal which may be differentiated from the pilot signals transmitted by the other N1 antenna instances by its position in the timefrequency plane. These pilot signals enable the OTFS receiver to separately measure each channel and the coupling between each antenna instance. Using this information the receiver essentially initializes the filters present within the equalizer 4000 such that convergence can be achieved more rapidly. In one embodiment an adaptive process is utilized to refine the inverted channel or filter used in separating the received signal energy into different timefrequencyspace planes. Thus, the coupling channel between each transmit and receive antenna instance may be measured, the representation of the measured channel inverted, and that inverted channel representation used to separate the received signal energy into separate and distinct timefrequency planes of information.
As noted above, the channel models associated with known conventional communication systems, such as OFDMbased systems, are onedimensional in nature. As such, these models are incapable of accurately taking into consideration all of the twodimensional (i.e., timebased and frequencybased) characteristics of the channel, and are limited to providing an estimate of only one such characteristic. Moreover, such onedimensional channel models change rapidly relative to the time scale of modern communication systems, and thus inversion of the applicable channel representation becomes very difficult, if possible at all.
The stationary twodimensional timefrequency channel models described herein also enable OFTS systems to effectively implement crosspolarization cancellation. Consider the case in which a transmit antenna instance associated with an OFTS transceiver is configured for horizontallypolarized transmission and a nearby receive antenna of the OFTS transceiver is configured to receive verticallypolarized signal energy. Unfortunately, reflectors proximate either the transmit or receive antenna may reflect and crosspolarize some of the transmitted horizontallypolarized energy from the transmit antenna, some of which may be directed to the receive antenna as a verticallypolarized reflection. It is believed that a twodimensional channel model of the type disclosed herein is needed in order to decouple and cancel this crosspolarized reflection from the energy otherwise intended for the receive antenna.
Similarly, full duplex communication carried out on the same channel requires echo cancellation sufficiently robust to substantially remove the influence of a transmitter on a nearby receiver. Again, such echo cancellation is believed to require, particularly in the case of moving reflectors, an accurate twodimensional representation of at least the echo channel in order to permit the representation to be appropriately inverted.
As discussed above, embodiments of the OTFS method may involve generating a twodimensional matrix by spreading a twodimensional input data matrix. In addition, time/frequency tiling may be utilized in transport of the twodimensional matrix across a channel. In this approach each matrix column may be tiled as a function of time; that is, each column element occupies a short symbol time slice utilizing the full available transmission bandwidth, with time gaps optionally interposed between subsequent columns. Alternatively, the matrix columns may be tiled and transported as a function of frequency; that is, each element of the column occupies a frequency bin for a longer period of time, with time gaps optionally interposed between subsequent columns.
In other embodiments a spreading kernel may be used to effect spreading of the input data matrix. In this case twodimensional spreading may be achieved through, for example, a twodimensional cyclic convolution with a spreading kernel, a convolution implemented using a twodimensional FFT, multiplication with the twodimensional DFT of the spreading kernel, followed by a twodimensional inverse Fourier transform. A wide variety of spreading kernels may be utilized; however, the twodimensional DFT of the selected kernel should lack any zeroes so as to avoid division by zero during the dispreading process. Moreover, spreading may also be achieved using alternate methods of convolutions, transforms and permutations. Masking (i.e., element by element multiplication) may also be utilized as long as each operation in invertible.
Attention is now directed to
As shown in
A receiver 4420 of the first OTFS transceiver 4400 includes first and second inverse timefrequency tiling elements 4424, 4426 configured to effect an inverse of the tiling operation performed by the timefrequency tiling elements 4412 and 4414. A twodimensional IFFT block 4428 and despreading block 4430 are configured to perform the inverse of the spreading operation performed by the twodimensional spreading block 4408 and the FFT block 4410. The received data is then converted using an FFT block 4434 prior to being equalized by a timefrequencyspace decision feedforward/feedback analyzer block 4438. The equalized data is then converted using an IFFT block 4440.
Turning now to
As shown in
A receiver 4470 of the second OTFS transceiver 4450 includes a serial arrangement of inverse timefrequency tiling elements 4474, 4476, 4478 configured to effect an inverse of the tiling operation performed by the timefrequency tiling elements 4462, 4464 and 4466. A multiplier 4480 is configured to multiply the output produced by the inverse timefrequency tiling elements 4474, 4476 and 4478 by an inverse mask. Next, an IFFT block 4482 converts the output of the multiplier 4480 and provides the result to a timefrequencyspace decision feedforward/feedback analyzer block 4488. The equalized data is then converted by an IFFT block 4492.
Attention is now directed to
Referring to
The mesh network 5000 comprises a plurality of OTFS wireless mesh nodes 5020 operative to provide wireless communication coverage to fixed or mobile devices within areas of high demand which are generally outside of the range of the coverage areas 5008. For the reasons discussed above, each OTFS wireless mesh node 5020 may be configured for full duplex wireless communication with other such mesh nodes 5020 over the same frequency band. This full duplex wireless communication over the same frequency band is represented in
Turning now to
In one embodiment mesh spatial gain may be achieved by using neighboring mesh nodes 5120 to support the simultaneous parallel transmission of streams of information using an identical frequency band over a single point to point link. This approach may improve signal transmission gain by using neighboring nodes 5120 to effectively create a distributed transmit source, thereby achieving gain through spatial signal separation.
Some embodiments of the systems and methods described herein may include computer software and/or computer hardware/software combinations configured to implement one or more processes or functions associated with the methods such as those described above and/or in the related applications. These embodiments may be in the form of modules implementing functionality in software and/or hardware software combinations. Embodiments may also take the form of a computer storage product with a computerreadable medium having computer code thereon for performing various computerimplemented operations, such as operations related to functionality as describe herein. The media and computer code may be those specially designed and constructed for the purposes of the claimed systems and methods, or they may be of the kind well known and available to those having skill in the computer software arts, or they may be a combination of both.
Examples of computerreadable media within the spirit and scope of this disclosure include, but are not limited to: magnetic media such as hard disks; optical media such as CDROMs, DVDs and holographic devices; magnetooptical media; and hardware devices that are specially configured to store and execute program code, such as programmable microcontrollers, applicationspecific integrated circuits (“ASICs”), programmable logic devices (“PLDs”) and ROM and RAM devices. Examples of computer code may include machine code, such as produced by a compiler, and files containing higherlevel code that are executed by a computer using an interpreter. Computer code may be comprised of one or more modules executing a particular process or processes to provide useful results, and the modules may communicate with one another via means known in the art. For example, some embodiments of systems described herein may be implemented using assembly language, Java, C, C#, C++, or other programming languages and software development tools as are known in the art. Other embodiments of the described systems may be implemented in hardwired circuitry in place of, or in combination with, machineexecutable software instructions.
The foregoing description, for purposes of explanation, used specific nomenclature to provide a thorough understanding of the claimed systems and methods. However, it will be apparent to one skilled in the art that specific details are not required in order to practice the systems and methods described herein. Thus, the foregoing descriptions of specific embodiments of the described systems and methods are presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the claims to the precise forms disclosed; obviously, many modifications and variations are possible in view of the above teachings. The embodiments were chosen and described in order to best explain the principles of the described systems and methods and their practical applications, they thereby enable others skilled in the art to best utilize the described systems and methods and various embodiments with various modifications as are suited to the particular use contemplated. It is intended that the following claims and their equivalents define the scope of the systems and methods described herein.