US 2019/0042536 A1 · Subspace-Constrained Partial Update Method For High-Dimensional Adaptive Processing Systems

Abstract

A method is explained for any adaptive processor processing digital signals by adjusting signal weights on digital signal(s) it handles, to optimize adaptation criteria responsive to a functional purpose or externalities (transient, temporary, situational, and even permanent) of that processor. Adaptation criteria for the adaptive algorithm may be any combination of a signal or parameter estimation, and measured quality(ies). This method performs a linear transformation adapting parameters from M to (M1+L) dimensions in each adaptation event, such that M1 weights are updated without constraints and M0=M−M1 weights are forced by soft constraints into an L-dimensional subspace they spanned at the beginning of the adaptation period. The same dimensionality reduction, using the same linear transformation, is applied to the input data. The reduced-dimensionality weights are then adapted using the identical optimization strategy employed by the processor, except with input data that has also been reduced in dimensionality.

Claims12

Dependent claims

1) A method of implementing partial updates in an adaptive processor that adjusts weights to optimize an adaption criterion in a signal estimation or parameter estimation algorithm comprising, during each of a plurality of successive cycles:

a) generating held-set indices and update-set indices;
b) separating input data into a held-set data and update-set data;
c) extracting held-set weights from full weights stored in memory;
d) computing enhanced data from the held-set weights, and held-set data, and update-set data;
e) computing reduced dimensionality weights using an optimization strategy of arbitrary type and structure, the enhanced data operated on by the optimization strategy, the optimization strategy thereby effecting the digital signal processing on the adaptive processor comprising any of the set of:
reduction or elimination of misadjustment effects, reduction in complexity and therefore cost of processing, improving efficiency, and providing an inherently more stable approach on a block-by-block basis without introducing hard constraints that lead to jitter;
f) generating full weights;
and
g) storing the full weights in memory.

6 more elements

5) A method of implementing partial updates in an adaptive processor that adjusts weights to optimize an adaption criterion in a signal estimation or parameter estimation algorithm comprising, during each of a plurality of successive cycles:

a) updating a dimensionality reduction strategy;
b) computing a dimensionality reduction transformation;
c) computing enhanced data using the dimensionality reduction transformation;
d) computing reduced dimensionality weights using an optimization strategy of arbitrary type and structure, the enhanced data operated on by the optimization strategy, the optimization strategy thereby effecting the digital signal processing on the adaptive processor comprising any of the set of:
reduction or elimination of misadjustment effects, reduction in complexity and therefore cost of processing, improving efficiency, and providing an inherently more stable approach on a block-by-block basis without introducing hard constraints that lead to jitter;
e) generating full weights;
and
f) storing the full weights in memory.

5 more elements

9) A method of implementing partial updates in an adaptive processor that adjusts weights to optimize an adaption criterion in a signal estimation or parameter estimation algorithm comprising, during each of a plurality of successive cycles:

a) generating held-set indices and update-set indices;
b) separating input data into a held-set data and update-set data;
c) extracting held-set weights from full weights stored in memory;
d) combining the held-set data and held-set weights to form held-set output data;
e) computing enhanced data from the update-set data and the held-set output data;
f) optimizing the enhanced data using an optimization strategy of arbitrary type and structure to form reduced dimensionality weights, the enhanced data operated on by the optimization strategy, thereby effecting the digital signal processing on the adaptive processor comprising any of the set of:
reduction or elimination of misadjustment effects, reduction in complexity and therefore cost of processing, improving efficiency, and providing an inherently more stable approach on a block-by-block basis without introducing hard constraints that lead to jitter;
g) generating full weights;
and
h) storing the full weights in memory.

7 more elements

Dependencies are inferred from the claim text.

Subspace-Constrained Partial Update Method For High-Dimensional Adaptive Processing Systems

Life of this patent

Subspace-Constrained Partial Update Method For High-Dimensional Adaptive Processing Systems

Life of this patent

Abstract

Claims12