Decision feedback analyzer with filter compensation

• US 6,751,255 B1
• Filed: 03/09/2000
• Issued: 06/15/2004
• Est. Priority Date: 03/09/2000
• Status: Expired due to Term

• Current Assignee: RPX Corporation
• Original Assignee: Orckit Communications Ltd. (Orckit-Corrigent )
• Inventors: Reuven, Ilan, Shahar Doron, Ayelet
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): ODOM, CURTIS B
• Application Number:
  US09/521,495
• Time in Patent Office: 1,559 Days
• Field of Search: 375/229, 375/285, 375/233, 375/232, 375/214, 375/230, 348/607
• US Class Current: 375/233
• CPC Class Codes: H03H 21/0012 Digital adaptive filters
• CPC Class Codes: H04L 2025/0349 as a feedback filter
• CPC Class Codes: H04L 2025/03503 as a combination of feedback and prediction filters
• CPC Class Codes: H04L 2025/03617 Time recursive algorithms H04L2025/03643 takes precedence
• CPC Class Codes: H04L 25/03057 with a recursive structure H04L25/03031 takes precedence
• CPC Class Codes: H04L 25/03343 Arrangements at the transmitter end
• CPC Class Codes: H04L 25/4975 Correlative coding using Tomlinson precoding, Harashima precoding, Trellis precoding or GPRS

Abstract

A digital data receiver, for receiving a signal that is transmitted over a channel having a channel response and is filtered with a predetermined filter response. The receiver includes a decision device, coupled to receive a sequence of equalized samples and to generate, responsive thereto, a corresponding sequence of estimated symbols. A feed-forward equalizer receives a sequence of input samples of the signal and generates responsive to the input samples, forward-equalized samples. A decision feedback equalizer (DFE), which receives the sequence of estimated symbols, includes an adaptive element, which processes the estimated symbols responsive to the channel response, and a fixed element, which processes the estimated symbols responsive to the predetermined filter response. The DFE generates feedback-equalized samples by combined operation of the adaptive and fixed elements. An adder combines the forward-equalized samples with the feedback-equalized samples so as to provide the equalized samples to the decision device.

Claims

1. A digital data receiver, for receiving a signal that is transmitted over a channel having a channel response and is filtered with a predetermined filter response, the receiver comprising:

a decision device, coupled to receive a sequence of equalized samples and to generate, responsive thereto, a corresponding sequence of estimated symbols;

a feed-forward equalizer, coupled to receive a sequence of input samples of the signal and to generate, responsive to the input samples, forward-equalized samples;

a decision feedback equalizer (DFE), coupled to receive the sequence of estimated symbols, and comprising;

an adaptive element, operative to process the estimated symbols responsive to the channel response; and

a fixed element, operative to process the estimated symbols responsive to the predetermined filter response,

the DFE generating feedback-equalized samples by combined operation of the adaptive and fixed elements; and

an adder, coupled to combine the forward-equalized samples with the feedback-equalized samples so as to provide the equalized samples to the decision device.

2. A receiver according to claim 1, and comprising a filter, which is coupled to apply the predetermined filter response to the signal.

3. A receiver according to claim 1, wherein the predetermined filter response is applied by a transmitter, which transmits the signal over the channel.

4. A receiver according to claim 1, wherein the predetermined filter response comprises a notch filter response.

5. A receiver according to claim 1, wherein the predetermined filter response is applied by a digital filter having an infinite impulse response, and wherein the fixed element of the DFE comprises an infinite impulse response filter element.

6. A receiver according to claim 1, wherein the adaptive element comprises:

a decision feedback filter having adaptive coefficients; and

an adaptation circuit, operative to determine the adaptive coefficients responsive to the predetermined filter response.

7. A receiver according to claim 6, wherein the adaptation circuit determines the adaptive coefficients based on filtering the estimated symbols using the predetermined filter response.

8. A receiver according to claim 7, wherein the adaptation circuit determines the adaptive coefficients using a least-mean-square method.

9. A receiver according to claim 1, wherein the adaptive element comprises:

a decision feedback filter having adaptive coefficients; and

an adaptation circuit, operative to determine the adaptive coefficients using least-mean-square method.

10. A receiver according to claim 1, wherein the predetermined filter response is described by a frequency response F(z), and

wherein the fixed element of the DFE comprises:

a first fixed element, having a frequency response substantially equal to F(z), which operates on the estimated symbols in conjunction with the adaptive element to generate a first output; and

a second fixed element, having a frequency response substantially equal to (F(z)−

1)/z<sup>−

1</sup>, which operates on the estimated symbols to generate a second output, and

wherein the DFE comprises an adder, which sums the first and second outputs to generate the feedback-equalized samples.

11. A receiver according to claim 1, wherein the adaptive element comprises a decision feedback filter having adaptive coefficients, which are determined responsive to the channel response in an initial stage of operation of receiving the signal from a transmitter, and

wherein in a subsequent stage of receiving the signal, the adaptive coefficients are conveyed to a precoder in the transmitter that is operative to precode the signal using the coefficients.

12. A receiver according to claim 11, wherein in the subsequent stage of receiving the signal, the adaptive element of the DFE is substantially inoperative, while the fixed element continues to process the estimated symbols responsive to the predetermined filter response.

13. A method for decision feedback equalization, comprising:

generating a sequence of estimated symbols based on a corresponding sequence of equalized samples, which are derived from a signal that is transmitted over a channel having a channel response and is filtered with a predetermined filter response;

processing the estimated symbols with adaptive processing parameters responsive to the channel response so as to generate adaptive feedback samples;

processing the estimated symbols with fixed processing parameters responsive to the predetermined filter response so as to generate fixed feedback samples; and

combining the adaptive feedback samples with the fixed feedback samples so as to generate feedback-equalized samples, for use in deriving the equalized samples.

14. A method according to claim 13, wherein generating the sequence of estimated symbols comprises receiving the signal over the channel and filtering the received signal with the predetermined filter response.

15. A method according to claim 13, wherein generating the sequence of estimated symbols comprises receiving the signal, which was filtered with the predetermined filter response before being transmitted over the channel.

16. A method according to claim 13, wherein the predetermined filter response comprises a notch filter response.

17. A method according to claim 13, wherein the predetermined filter response comprises an infinite impulse response, and wherein processing the estimated symbols with the fixed processing parameters comprises applying an infinite impulse response filter to the estimated symbols.

18. A method according to claim 13, wherein combining the adaptive and fixed feedback samples so as to generate the equalized samples comprises combining the adaptive and fixed feedback samples with forward-equalized samples derived from the signal, so as to generate the equalized samples.

19. A method according to claim 13, wherein processing the estimated symbols with the adaptive processing parameters comprises applying a decision feedback equalizer having adaptive coefficients to the estimated symbols, and determining the adaptive coefficients responsive to the predetermined filter response.

20. A method according to claim 19, wherein determining the adaptive coefficients comprises determining the coefficients using a least-mean-square method.

21. A method according to claim 13, wherein the predetermined filter response is described by a frequency response F(z), and

wherein processing the estimated symbols with the adaptive processing parameters comprises filtering the estimated symbols with a frequency response substantially equal to F(z) in conjunction with applying the adaptive processing parameters, so as to generate the adaptive feedback samples, and

wherein processing the estimated symbols with the fixed processing parameters comprises filtering the estimated symbols with a frequency response substantially equal to (F(z)− 1)/z<sup>−

1</sup>, so as to generate the fixed feedback samples, and

wherein combining the adaptive feedback samples with the fixed feedback samples comprises summing the adaptive and fixed feedback samples.

22. A method according to claim 13, and comprising:

receiving the signal over the channel from a transmitter having a precoder, wherein processing the estimated symbols with the adaptive processing parameters comprises applying a decision feedback equalizer having adaptive coefficients to the estimated symbols, and determining the adaptive coefficients responsive to the channel response in an initial stage of receiving the signal from the transmitter; and

in a subsequent stage of receiving the signal, conveying the adaptive coefficients to the precoder, wherein the precoder is operative to precode the signal using the coefficients.

23. A method according to claim 22, wherein processing the estimated symbols with the adaptive processing parameters comprises ceasing to process the estimated symbols with the adaptive processing parameters in the subsequent stage of receiving the signal, while continuing to process the estimated symbols with the fixed processing parameters responsive to the predetermined filter response.