Wideband digital equalizers for subscriber loops

• US 5,293,402 A
• Filed: 05/02/1991
• Issued: 03/08/1994
• Est. Priority Date: 05/02/1991
• Status: Expired due to Term

• Current Assignee: TTI Inventions A LLC (Intellectual Ventures LLC)
• Original Assignee: Bell Communications Research, Inc. (Telefonaktiebolaget LM Ericsson)
• Inventors: Honig, Michal L., Crespo, Pedro M.
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): TSE, YOUNG TOI
• Application Number:
  US07/694,819
• Time in Patent Office: 1,041 Days
• Field of Search: 375/12, 375/13, 375/14, 375/101, 375/103, 333/18, 333/28 R, 364/724.2, 364/724.17, 370/32.1
• US Class Current: 375/233
• CPC Class Codes: H04B 3/04 Control of transmission; Equalising
• CPC Class Codes: H04L 25/03057 with a recursive structure H04L25/03031 takes precedence

Abstract

A digital decision feedback equalizer utilizes a finite impulse response (FIR) filter and an infinite impulse response (IIR) filter in parallel in the feedback path. The FIR filter compensates for the rapidly changing region of the channel impulse response immediately following the cursor while the IIR filter compensates for the gradually decaying tail of the impulse response of the channel. Modifications of a sequential identification algorithm are used to adapt the filter sections. In particular, in one modification, the inputs to both the poles and zeros sections of the IIR filter are prefiltered by an all-pole filter, where the poles of this filter are the current pole estimates. In another modification, the order of the zeros section of the IIR filter and the all-pole filter are reversed, and the zeros section is adapted using the least mean squares algorithm. A recursive least squares algorithm with exponential fading can be used to adapt the poles and/or zeros section of the IIR filter.

Claims

1. A decision filter equalizer for receiving digital pulse transmissions comprising

a finite impulse response filter in the feedback path of said equalizer,

an infinite impulse response filter in said feedback path in parallel with said finite impulse response filter,

said infinite impulse response filter having a zeros section and a poles section, and

means utilizing a sequential identification algorithm for adapting said finite impulse response filter,

the zeros section of said infinite impulse response filter preceding the poles section in said feedback path of said infinite impulse response filter.

2. A decision feedback equalizer for receiving digital pulse transmission comprising

a finite impulse response filter in the feedback path of said equalizer,

an infinite impulse response filter in said feedback path in parallel with said finite impulse response filter,

said infinite impulse response filter having a zeros section and a poles section, and

adapting means utilizing a recursive least squares adaptation algorithm with exponentially decaying memory for adapting said infinite impulse response filter.

3. A decision feedback equalizer for receiving digital pulse transmissions comprising

a finite impulse response filter in the feedback path of said equalizer,

an infinite impulse response filter in said feedback path in parallel with said finite impulse response filter,

said infinite impulse response filter having a zeros section and a poles section,

the zeros section of said infinite impulse response filter preceding the poles section of said infinite impulse response filter, and

adapting means utilizing a sequential identification algorithm for adapting said infinite impulse response filter, ##EQU14## where β

is the step size,

i is the sample index

k is the tap index of said infinite impulse response filter,

c_k is the kth tap weight for the numerator C or zeros section of the transfer function of said infinite response filter,

d_k is the kth tap weight for the denominator D or poles section of the transfer function of said infinite impulse response filter,

τ

is the total delay, in intertap delay units, of said finite impulse response filter and (τ

-1) is the number of taps in said finite impulse response filter,

y(i) is the difference between the input signal to said equalizer and the output of said finite impulse response filter, and

s(i) is the ith output sample from said equalizer.

4. A decision feedback equalizer for receiving digital pulse transmissions comprising

a finite impulse response filter in the feedback path of said equalizer,

an infinite impulse filter in said feedback path, in parallel with said finite impulse response filter,

said infinite impulse response filter having a zeros section and a poles section,

the zeros section of said infinite impulse response filter preceding the poles section of said infinite impulse response filter, and

adapting means utilizing a recursive least squares adaptation algorithm with exponentially decaying memory for adapting said infinite impulse response filter,

said adapting means carrying out the following algorithm:

Initialize;

##EQU15## δ

is a small constant, I is the identity matrix,

C(i,q) is the tap weight of the poles section of the infinite impulse response filter at sample i,

g^c is the output of the poles section of said infinite impulse response filter in response to the input s,

g^d is the output of the poles section of said infinite impulse response filter in response to the input (y-s),

e is the error estimate used to adapt said equalizer,

R is a sample correlation matrix of partially processed signals, and

w is a constant slightly less than one.