MLSE decoding of PRS type inter-bin interference in receiver-end windowed DMT system

• US 7,058,141 B1
• Filed: 06/02/2000
• Issued: 06/06/2006
• Est. Priority Date: 06/02/2000
• Status: Expired due to Term

• Current Assignee: NEC Corporation
• Original Assignee: NEC USA Inc (NEC Corporation)
• Inventors: Nedic, Slobodan
• Primary Examiner(s): BURD, KEVIN MICHAEL
• Application Number:
  US09/586,441
• Time in Patent Office: 2,195 Days
• Field of Search: 375/341, 375/259, 375/346, 375/260, 375/316, 375/340, 714794-796
• US Class Current: 375/341
• CPC Class Codes: H04L 1/0054 Maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
• CPC Class Codes: H04L 2025/03414 Multicarrier
• CPC Class Codes: H04L 2025/03445 Time domain
• CPC Class Codes: H04L 2025/03522 Frequency domain
• CPC Class Codes: H04L 25/03197 methods of calculation involving metrics
• CPC Class Codes: H04L 25/497 by correlative coding, e.g. partial response coding or echo modulation coding transmitters and receivers for partial response systems transversal equalizers at the transmitter end H04L25/03343

Abstract

An improved method of decoding the Partial-Response Signaling type inter-bin distortion in receiver-end windowed DMT system is disclosed. The method exploits the Maximum Likelihood Sequence Estimation, which yields certain improvements over the conventional DFE-type decoding of PRS signals. Simulation results are shown for the case wherein AWGN is substantially the only impairment. Since in this case the noise samples at the neighboring frequency bin outputs bear correlation of the same type as the correlation introduced among data symbol samples, a trellis imbedded noise prediction/cancellation approach has been devised. It has been shown that performance degradation can be reduced from almost 8 dB to just 1 dB, compared to performance figures with no windowing. Possible applications of this method are for xDSL and wireless OFDM systems.

Claims

1. A receiver circuit for providing a decoded output from a received discrete multi-tone modulated input signal, the input signal being received from a communication channel having noise thereon, the input signal comprising digital data, the circuit comprising:

a first stage having a frequency response for applying a discrete Fourier transform to the input signal; and

a second stage for receiving an output signal from said first stage and per-bin decoding said signal based on a maximum likelihood sequence estimation (MLSE) algorithm so as to recover said digital data;

wherein said maximum likelihood sequence estimation (MLSE) algorithm includes a calculation of state metrics and is modified to include subtraction of estimated noise samples using noise prediction coefficients.

2. The circuit of claim 1, wherein said noise prediction coefficients are the same for all bins.

3. The circuit of claim 1, wherein at least two bins or groups of bins have different respective noise prediction coefficients.

4. The circuit of claim 1, wherein said modified MLSE algorithm performs cancellation of noise received with said input signal from said communication channel.

5. The circuit of claim 4, wherein said noise is additive noise, additionally colored by a windowing procedure.

6. The circuit of claim 5, wherein said windowing procedure is time domain windowing.

7. The circuit of claim 5, wherein said windowing procedure is frequency domain windowing.

8. A method in a receiver circuit for providing a decoded output from a received discrete multi-tone modulated input signal, the input signal being received from a communication channel having noise thereon, the input signal comprising digital data, the method comprising the steps of:

applying a discrete Fourier transform to the input signal; and

then

per-bin decoding said signal based on a maximum likelihood sequence estimation algorithm so as to recover said digital data;

wherein said maximum likelihood sequence estimation (MLSE) algorithm includes a calculation of state metrics and is modified to include subtraction of estimated noise samples using noise prediction coefficients.

9. The method of claim 8, wherein said noise prediction coefficients are the same for all bins.

10. The method of claim 8, wherein at least two bins or groups of bins have different respective noise prediction coefficients.

11. The method of claim 8, further comprising the step of utilizing said modified MLSE algorithm to perform cancellation of noise received with said input signal from said communication channel.

12. The method of claim 11, wherein said noise is additive noise, additionally colored by a windowing procedure.

13. The method of claim 12, wherein said windowing procedure applies time domain windowing.

14. The method of claim 12, wherein said windowing procedure applies frequency domain windowing.