System and method of estimating equalizer performance in the presence of channel mismatch

• US 5,513,214 A
• Filed: 07/27/1995
• Issued: 04/30/1996
• Est. Priority Date: 03/05/1992
• Status: Expired due to Term

• Current Assignee: Lockheed Martin Corporation (Martin Marietta Corporation)
• Original Assignee: Loral Federal Systems Company (Martin Marietta Corporation)
• Inventors: Gozzo, Francesco
• Primary Examiner(s): Safourek, Benedict V.
• Assistant Examiner(s): PHAN, HAI
• Application Number:
  US08/514,356
• Time in Patent Office: 278 Days
• Field of Search: 375/229-236, 375/346, 375/350, 364/724.19, 364/724.2, 333/18, 333/28 R
• US Class Current: 375/232
• CPC Class Codes: H03H 21/0012 Digital adaptive filters
• CPC Class Codes: H04L 2025/03566 between different tapped delay line structures
• CPC Class Codes: H04L 25/03057 with a recursive structure H04L25/03031 takes precedence

Abstract

Linear and decision feedback equalizers which are communication receivers used in computer systems having an ISI receiver system and analysis paradigm to provide the ability of estimating the performance of these receivers under mismatched channel conditions. The mean square error (MSE) performance of linear and decision feedback equalizers in the presence of arbitrary channel mismatch is presented. A generic equalizer which has feed-forward and feedback taps which are optimized for a mean square error (MSE) criterion. The ability to rapidly predict MSE performance also leads to the real-time adaptation of the equalizer complexity to improve MSE performance, optimal truncation of channel impulse response based on an MSE criterion, and assessing the impact of assuming white noise in a colored noise environment. With the system and method there is a determination of the spectral characteristics of the communication channel during training and/or decoding. These are either known in advance via models and/or experimentation, or they are estimated and/or measured on-line on a periodic schedule. Once this data is obtained, the MSE estimation paradigm is used to predict the MSE performance of the linear or DFE receiver. Among the areas of application is the area of optical and wireless communication systems, such as mobile cellular systems, satellite systems, and wireless indoor networks. Since the physical channel in wireless systems is susceptible to many variations, the receiver is usually operating with inaccurate channel estimates.

Claims

1. A data communication system comprising:

a linear equalizer;

a receiving system for use with the linear equalizer for providing a mean square error (MSE) of a performance of the linear equalizer under arbitrary channel mismatch conditions, the receiving system including;

receiving means for receiving data sent to the communication system, the receiving means including channel means mismatched with the linear equalizer;

the channel means coupled to the linear equalizer for transmitting to the linear equalizer the data sent to the communication system, the channel means contributing filter and noise components to the data sent to the communication system which causes said arbitrary channel mismatch conditions;

the linear equalizer including;

input means coupled to the channel means for receiving the data sent to the communication system transmitted by the channel means and having the filter and noise components contributed by the channel means, the filter components described by |H_t (e^jw)|² during a training sequence for the linear equalizer and by |H(e^jw)|² during decoding, the noise components described by N_t (e^jw) during the training sequence and N(e^jw) during the decoding; and

means for outputting an approximation of the data sent to the communication system without the filter and noise components;

the MSE determined according to ##EQU8##

2. A data communication system as claimed in claim 1, wherein the noise components include white gaussian noise and the MSE is determined by setting N(e^jw)=N₀, N_t (e^jw)=N₀ +δ

_n, and H_t (e^jw)=H(e^jw), where N₀ is a noise spectrum during the decoding, and δ

_n represents a difference in the noise spectrum during the training.

3. A data communication system as claimed in claim 1, wherein the filter and noise components are pre-determined from estimated time-domain data or from channel models.

4. A data communication system as claimed in claim 1, wherein the MSE is used to determine optimal truncation of channel impulse responses.

5. A data communication system as claimed in claim 1, wherein the MSE is used to determine a magnitude of error caused by ignoring white noise.

6. A data communication system as claimed in claim 1, wherein the MSE is used to determine design strategies in multi-media and storage applications.

7. A data communication system as claimed in claim 1, wherein the linear equalizer is a Kalman type equalizer or a derivative of the Kalman type equalizer.

8. A data communication system as claimed in claim 1, wherein the data sent to the communication system includes non-binary signalling.

9. A data communication system as claimed in claim 1, wherein the MSE is used for determining an optimal number of feedforward taps in the linear equalizer.

10. A data communication system comprising:

a decision feedback equalizer.(DFE);

a receiving system for-use with the DFE for providing a mean square error (MSE) of a performance of the DFE under arbitrary channel mismatch conditions, the receiving system including;

receiving means for receiving data sent to the communication system, the receiving means including channel means mismatched with the DFE;

the channel means coupled to the DFE for transmitting to the DFE the data sent to the communication system, the channel means contributing filter and noise components to the data sent to the communication system which causes said arbitrary channel mismatch conditions;

the DFE including;

input means coupled to the channel means for receiving the data sent to the communication system transmitted by the channel means and having the filter and noise components contributed by the channel means, the filter components described by |H_t (e^jw)|² during a training sequence for the DFE and by |H(e^jw)|² during decoding, the noise components described by N_t (e^jw) during the training sequence and N(e^jw) during the decoding; and

means for outputting an approximation of the data sent to the communication system without the filter and noise components;

the MSE determined according to ##EQU9##

11. A data communication system as claimed in claim 10, wherein the noise components include white gaussian noise and the MSE is determined by setting N(e^jw)=N₀, N_t (e^jw)=N₀ +δ

_n, and H_t (e^jw)=H(e^jw), where N₀ is a noise spectrum during the decoding, and δ

_n represents a difference in the noise spectrum during the training.

12. A data communication system as claimed in claim 10, wherein the filter and noise components are pre-determined from estimated time-domain data or from channel models.

13. A data communication system as claimed in claim 10, wherein the MSE is used to determine optimal truncation of channel impulse responses.

14. A data communication system as claimed in claim 10, wherein the MSE is used to determine a magnitude of error caused by ignoring white noise.

15. A data communication system as claimed in claim 10, wherein the MSE is used to determine design strategies in multi-media and storage applications.

16. A data communication system as claimed in claim 10, wherein the DFE is a Kalman type equalizer or a derivative of the Kalman type equalizer.

17. A data communication system as claimed in claim 10, wherein the data sent to the communication system includes non-binary signalling.

18. A data communication system as claimed in claim 10, wherein the MSE is used for determining an optimal number of feedforward taps in the DFE.

19. A data retrieval system comprising:

a storage system having a read/write head for retrieving and receiving data from the storage system;

a receiving system for providing a mean square error (MSE) of a performance of the storage system, the receiving system including;

a data channel coupled to the read/write head for carrying the data retrieved from the storage system, the data channel contributing filter and noise components;

equalizer means coupled to the data channel for receiving the data retrieved from the storage system having the filter and noise components, and for eliminating the filter and noise components from the data retrieved from the storage system, the filter components described by |H_t (e^jw)|² during training of the receiving system and by |H(e^jw)|² during decoding, the noise components described by N_t (e^jw) during the training of the receiving system and N(e^jw) during the decoding, the MSE determined according to ##EQU10## if the equalizer means is a linear equalizer, and according to ##EQU11## if the equalizer means is a decision feedback equalizer.

20. A method for operating and evaluating an equalizer, comprising the steps of:

transmitting an information sequence to the equalizer over an input channel;

the equalizer receiving an input sequence comprising the information sequence, channel filter spectral data, and channel noise;

pre-determining and defining the channel filter spectral data according to |H_t (e^jw)|² ;

training the equalizer according to the pre-determined channel filter spectral data defined by |H_t (e^jw)|² ;

pre-determining and defining the channel noise according to N_t (e^jw);

training the equalizer according to the pre-determined channel noise defined by N_t (e^jw);

the equalizer outputting an estimate of the information sequence in response to the input sequence received over the input channel according to channel filter spectral data defined by |H(e^jw)|² and channel noise defined by N(e^jw);

determining a mean square error of the equalizer according to ##EQU12## the equalizer is a linear equalizer; and

determining a mean square error of the equalizer according to ##EQU13## if the equalizer is a decision feedback equalizer.

21. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for evaluating a performance of a receiver by providing a mean square error of the receiver'"'"'s output, said method steps comprising:

retrieving receiver data which describes a type of the receiver;

retrieving filter spectral data which describes characteristics of a receiver input channel filter;

defining the filter spectral data according to |H_t (e^jw)|² during training of the receiver and according to |H(e^jw)|² during decoding;

retrieving noise spectral data which describes noise characteristics of the receiver input channel;

defining the noise spectral data according to N_t (e^jw) during the training of the receiver and according to N(e^jw) during the decoding;

processing the filter and the noise spectral data to form intermediate variables;

determining the MSE of the receiver according to ##EQU14## if the receiver data indicates that the receiver is a linear equalizer; and

determining the MSE of the receiver according to ##EQU15## if the receiver data indicates that the receiver is a decision feedback equalizer.

22. An article of manufacture for use in a computer system having a receiver for receiving transmitted data;

said article of manufacture comprising a computer usable medium having computer readable program code means embodied in said medium, said program code means for causing a computer to effect;

retrieving receiver data which describes a type of the receiver;

retrieving filter spectral data which describes characteristics of a receiver input channel filter;

defining the filter spectral data according to |H_t (e^jw)|² during training of the receiver and according to |H(e^jw)|² during decoding;

retrieving noise spectral data which describes noise characteristics of the receiver input channel;

defining the noise spectral data according to N_t (e^jw) during the training of the receiver and according to N(e^jw) during the decoding;

processing the filter and the noise spectral data to form intermediate variables;

determining the MSE of the receiver according to ##EQU16## if the receiver data indicates that the receiver is a linear equalizer; and

determining the MSE of the receiver according to ##EQU17## receiver is a decision feedback equalizer.

23. A receiving system for use with an adaptive receiver for providing a mean square error value of a performance of the receiver, the receiver for providing an output signal x'"'"'_k 80 which is an estimate of a transmitted information sequence x_k 10, in response to an input signal r_k 60 received from a discrete-time equivalent channel which has a linear channel filter 20 having channel filter coefficients being coupled to the transmitted information sequence x_k and additive noise 40, the channel filter coefficients defined by |H_t (e^jw)|² during training of the receiver and by |H(e^jw)|² during receiver decoding, the additive noise defined by N_t (e^jw) during the training of the receiver and by N(e^jw) during the receiver decoding, the MSE defined by ##EQU18##