Low-complexity tightly-coupled integration filter for step detection in a sensor-assisted GNSS receiver

• US 8,645,062 B2
• Filed: 01/18/2013
• Issued: 02/04/2014
• Est. Priority Date: 09/24/2008
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Roh, June Chul, Waters, Deric Wayne, Hosur, Srinath, Dutta, Goutam
• Primary Examiner(s): Tarcza, Thomas
• Assistant Examiner(s): Tissot, Adam
• Application Number:
  US13/745,186
• Publication Number:
  US 20130131983A1
• Time in Patent Office: 382 Days
• Field of Search: 701/472, 701/468, 701/500, 701/501, 701/526, 702/1, 702/160, 702/127, 702/155, 702/158, 702/187, 702/189, 702/193, 340/944, 377/24.2, 377/1, 377/13, 377/24, 337/00, 731/78.R
• US Class Current: 701/472
• CPC Class Codes: G01C 21/1654 with electromagnetic compass
• CPC Class Codes: G01C 21/28 with correlation of data from several navigational instruments
• CPC Class Codes: G01C 22/006 Pedometers
• CPC Class Codes: G01S 19/47 the supplementary measurement being an inertial measurement, e.g. tightly coupled inertial
• CPC Class Codes: G06F 17/16 Matrix or vector computation , e.g. matrix-matrix or matrix-vector multiplication, matrix factorization matrix transposition G06F7/78

Abstract

Embodiments of the invention provide a step detection. An accelerometer measurement in the form of a multi-dimensional acceleration vector is obtained. The magnitude of the accelerometer measurement is filtered using a low pass filter. A threshold for a down-crossing is provided as is a threshold for an up-crossing. A step detection is triggered if the magnitude of the accelerometer measurement is greater than or equal to the threshold for an up-crossing.

Claims

1. A method of step detection comprising:

obtaining accelerometer measurement a(k), from an accelerometer, where a(k) is multi-dimensional acceleration vector at a k-th sample;

filtering the magnitude of said accelerometer measurement |a(k)| using a low-pass filter;

providing a threshold for a down-crossing (dc);

providing a threshold for an up-crossing (uc);

trigging a step detection if said magnitude of said accelerometer measurement |a(k)| is greater than or equal to said threshold for an up-crossing (uc); and

provide said step detection to a global navigation satellite system pedestrian dead reckoning GNSSPDR blending filter.

2. The method of claim 1, wherein trigging a step detection if said magnitude of said accelerometer measurement |a(k)| is less than or equal to said threshold for a down-crossing (dc).

3. The method of claim 1, further comprising transitioning to a static state if there is no up-crossing or down-crossing event for a predetermined time period.

4. The method of claim 1, further comprising transitioning to a down-crossing (dc) state if said magnitude of said accelerometer measurement |a(k)| is equal to or below said threshold for a down-crossing (dc).

5. The method of claim 1, further comprising:

providing a time window from time instance for the latest down-crossing; and

triggering a step detection when a transition to an up-crossing (uc) state from down-crossing (dc) state within said time window.

6. A method for pedestrian dead reckoning (PDR) data conversion for global navigation satellite system pedestrian dead reckoning (GNSSPDR) integration by converting PDR position data to user velocity measured at a plurality of time instances where GNSS position or velocity estimates are available from GNSS, said method comprising:

obtaining a heading and a step length for a previous step;

calculating a partial step using said heading and said step interval for said previous step;

obtaining a pedestrian dead reckoning (PDR) position at a first (k) GNSS clock p(t_k) where t_k is first clock, comprising;

adding said partial step to said PDR position, if a step detection state is not in Static;

keeping said position for said previous step, if said step detection state is in Static;

updating said position at a GNSS clock previous in time (k−

1) to said first GNSS clock (t<sub>k−

1</sub>) by interpolating with a plurality of measured step instances; and

taking a difference to obtain said user velocity by using