Adaptive Filter Resource Sharing for Faster Coefficient Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive filters face challenges in optimizing computational resource usage, leading to inefficient power consumption and increased costs, particularly in implementing the adaptive convergence algorithm, where filter coefficients are updated at suboptimal rates due to fixed resource allocation.
Innovation Solution
The adaptive filter employs a resource sharing mechanism where computational blocks are dynamically allocated to adjust filter coefficients, allowing only a subset of coefficients to be updated at a time, with a sharing factor that determines the frequency of updates, and introduces offset injection to enhance convergence rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If computational blocks are shared among multiple filter coefficients, then power consumption and implementation costs are reduced, but filter coefficient convergence rate deteriorates
Solution Approach 1:
The patent implements dynamic resource allocation where the sharing factor is adjustable based on operational conditions. The system can switch between different sharing factors (e.g., sharing factor of 1 for full resources, sharing factor of 2 for reduced resources) depending on whether speed or power consumption is the priority, making the resource sharing dynamic rather than fixed
Solution Approach 2:
The patent changes the parameter of resource allocation by introducing an adjustable sharing factor that determines how many filter coefficients share a single computational block. This parameter can be modified to optimize the balance between power consumption and convergence rate based on specific application requirements
2Ease of manufacture
If computational blocks are shared among multiple filter coefficients, then implementation costs are reduced, but filter coefficient convergence rate deteriorates
Solution Approach 1:
The patent merges the functionality of multiple computational blocks into a single shared block that serves multiple filter coefficients. By combining resources, the system reduces the total number of computational blocks required, thereby lowering implementation costs while maintaining acceptable performance through the sharing mechanism
Solution Approach 2:
The patent creates universal computational blocks that can serve multiple purposes by adjusting the sharing factor. A single computational block can dynamically serve different numbers of filter coefficients based on the configured sharing factor, making the system more versatile and cost-effective
3Use of energy by moving object
If a subset of filter coefficients is updated at a time with resource sharing, then power consumption is reduced, but convergence speed deteriorates
Solution Approach 1:
The patent implements periodic action by cycling through different subsets of filter coefficients for updates. Instead of updating all coefficients simultaneously, the system updates subsets in periodic cycles, reducing instantaneous power consumption while ensuring all coefficients eventually converge through repeated updates over time
Data Source
AI summary
The present application relates to an adaptive filter using manageable resource sharing and a method of operating the adaptive filter. The adaptive filter comprises a cluster controller configured for allocating each of several computational blocks to one of several clusters and a routing controller for configuring the routing of tapped delay signals by a routing logic to the respective cluster in accordance with an allocation of the tapped delay signals to the clusters. Each of computational blocks is configured for adjusting one filter coefficient, ci(n), in one cycle of an iterative procedure according to an adaptive convergence algorithm. The number of computational blocks is less than an order of the adaptive filter.