Adaptive Weight Updates With Subspace Constraints to Cut Misadjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current partial update methods for adaptive signal processors with large numbers of weights suffer from severe misadjustment and increased complexity, particularly when applied to high-order optimization functions, leading to reduced performance and instability in applications like phased array radar and smart grid networks.
Innovation Solution
The method involves a linear transformation of processor parameters from M-dimensions to (M1+L)-dimensions, allowing M1 weights to be updated without constraints and M0 weights to be subjected to soft constraints, enabling dimensionality reduction in both weights and input data, and using the same optimization strategy for adaptation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If partial update methods are applied to adaptive signal processors with large numbers of weights, then device complexity is reduced, but manufacturing precision deteriorates due to severe misadjustment
Solution Approach 1:
The weight vector is segmented into two distinct subsets: an update set of weights that are modified during adaptation, and a held set of weights that remain fixed. This segmentation allows the processor to reduce complexity by updating only a portion of the weights while maintaining stability through the held weights, thereby reducing misadjustment effects.
Solution Approach 2:
Instead of updating all weights in the adaptive signal processor, the method applies partial action by updating only a selected subset of weights (the update set) while holding the remaining weights constant. This partial update approach reduces computational complexity and memory requirements while minimizing misadjustment by maintaining the stability of the held weights.
2Loss of time
If conventional partial update methods are used, then adapt block size is reduced, but reliability deteriorates due to increased misadjustment and instability
Solution Approach 1:
The method changes the parameter configuration by selectively updating only certain weight parameters while holding others fixed. This parameter change strategy allows for smaller adapt block sizes and faster convergence while maintaining reliability through the stability provided by the held weights, which prevent the misadjustment and instability that would occur with more aggressive updates.
3Measurement precision
If high-order optimization functions are applied to adaptive processors, then measurement precision is improved, but device complexity increases significantly
Solution Approach 1:
The method segments the weight update process to apply high-order optimization only to a selected update set of weights, while holding the remaining weights fixed. This segmentation enables the use of computationally intensive high-order optimization functions to improve measurement precision without requiring the entire system to handle the full computational burden, thereby managing device complexity.
Data Source
AI summary
An adaptive processor implements partial updates when it adjusts weights to optimize adaptation criteria in signal estimation, parameter estimation, or data dimensionality reduction algorithms. The adaptive processor designates some of the weights to be update weights and the other weights to be held weights. Unconstrained updates are performed on the update weights, whereas updates to the set of held weights are performed within a reduced-dimensionality subspace. Updates to the held weights and the update weights employ adapt-path operations for tuning the adaptive processor to process signal data during or after tuning.