Reduced complexity recursive least square lattice structure adaptive filter by means of limited recursion of the backward and forward error prediction squares

• US 20060288066A1
• Filed: 04/07/2006
• Published: 12/21/2006
• Est. Priority Date: 06/20/2005
• Status: Active Grant

• Current Assignee: Google Technology Holdings LLC (Alphabet Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Barron, David L., Piket, James B., Iwai, Kyle K.
• Granted Patent: US 7,734,466 B2
• Time in Patent Office: Days
• US Class Current: 708/323
• CPC Class Codes: H03H 2021/0049 Recursive least squares algorithm
• CPC Class Codes: H03H 2021/007 Computation saving measures; Accelerating measures
• CPC Class Codes: H03H 2021/0074 Reduction of the update frequency
• CPC Class Codes: H03H 21/0014 Lattice filters
• CPC Class Codes: H03H 21/0043 Adaptive algorithms

Abstract

A method for reducing a computational complexity of an m-stage adaptive filter is provided by updating recursively forward and backward error prediction square terms for a first portion of a length of the adaptive filter, and keeping the updated forward and backward error prediction square terms constant for a second portion of the length of the adaptive filter.

Claims

1. A method for an m-stage recursive least squares lattice structure adaptive filter, the method comprising:

receiving an input signal at the adaptive filter;

recursively updating filter coefficients for each stage of the adaptive filter, including

reducing a number of updates of a forward error prediction squares and backward error prediction squares of the filter coefficients, and

holding the forward error prediction squares and the backward error prediction squares constant during the remainder of the act of updating the filter coefficients; and

producing a filtered signal at an output of the adaptive filter.

2. The method of claim 1 wherein updating filter coefficients comprises limiting updating of the forward error prediction squares and the backward error prediction squares to a predefined portion of filter length.

3. The method of claim 2 wherein limiting updating of the forward error prediction squares and the backward error prediction squares to a predefined portion of the filter length comprises updating the forward error prediction squares and the backward error prediction squares of approximately one-half of all filter stages.

4. The method of claim 1 wherein receiving an input signal comprises receiving digital data representative of a time-varying signal.

5. A method for reducing computational complexity of an m-stage adaptive filter, the method comprising:

receiving an input signal at an input;

filtering the input signal to produce a filtered signal, including

updating recursively a forward error prediction square term for a first portion of a length of the adaptive filter;

updating recursively a backward error prediction square term for the first portion of the length of the adaptive filter; and

keeping the updated forward and backward error prediction square terms constant for a second portion of the length of the adaptive filter; and

providing the filtered signal at an output.

6. The method of claim 5 wherein updating the forward error prediction square term comprises updating the forward error prediction square term for approximately one-half of the length of the adaptive filter and updating the backward error prediction square term comprises updating the backward error prediction square term for approximately one-half of the length of the adaptive filter.

7. The method of claim 5 wherein further comprising:

updating all filter coefficients up to a predefined number of taps of the adaptive filter for a first set of filter updates; and

holding constant the forward error prediction square term and the backward error prediction square term while updating the rest of the filter coefficients during a second set of filter updates.

8. The method of claim 7 further comprising:

during the second set of filter updates, using the constant forward error prediction square term to update a backward reflection coefficient; and

using the constant backward error prediction square term to update a forward reflection coefficient.

9. An adaptive filter comprising:

an interface to receive an input signal;

a processor operative in conjunction with stored data and instructions to recursively update filter coefficients for each stage of the adaptive filter while producing a filtered signal, the processor configured to

reduce a number of updates of the forward error prediction squares and backward error prediction squares of the filter coefficients, and

hold the forward error prediction squares and the backward error prediction squares constant during the remainder of the act of updating the filter coefficients; and

an interface to provide the filtered signal as an output signal.

10. The adaptive filter of claim 9 wherein the processor is operative to

update recursively the forward error prediction squares for a first portion of a length of the adaptive filter;

update recursively the backward error prediction squares for the first portion of the length of the adaptive filter; and

keep the updated forward error prediction squares and backward error prediction squares constant for a second portion of the length of the adaptive filter.

11. The adaptive filter of claim 10 wherein the processor is operative to

update all filter coefficients up to a predefined number of taps of the adaptive filter for a first set of filter updates;

and

hold constant the forward error prediction squares and the backward error prediction squares while updating the rest of the filter coefficients during a second set of filter updates.