Multi-convex combination adaptive filtering method based on maximum correlation entropy

Abstract

The invention provides a multi-convex combination adaptive filtering method based on maximum correlation entropy, the weight vector is updated by the maximum correlation entropy criterion, The L asynchronous adaptive filters are combined to modulate the output results in convex combinations, and the one of the L adaptive filters with different step sizes which has the largest step size ensures theconvergence rate of the multi-convex adaptive filters at the beginning of the filter, and then The other L-1 filters adjust the weight ratio according to the mixing coefficient at different time periods, thereby controlling the convergence speed and the steady state amount.. The invention not only improves the convergence speed, but also maintains a low misalignment amount, can more flexibly adjust the proportion of multiple step sizes, and improves the re-convergence ability and the tracking ability.

Description

technical field [0001] The invention belongs to the field of signal processing, and in particular relates to a multi-convex combination adaptive filtering method based on maximum correlation entropy. Background technique [0002] In recent years, adaptive filter has rapidly developed into an optimal filtering method. Adaptive filter is a filtering method developed on the basis of linear filtering such as Wiener filtering and Kalman filtering. Because of its stronger adaptability and better filtering performance, it is used in engineering practice, especially in information It has been widely used in processing technology. [0003] The research object of adaptive filtering is a system or information process with uncertainty. The "uncertainty" here means that the mathematical model of the researched information processing process and its environment is not completely determined. There are some unknowns and random factors involved. Any actual information process has differen...

AI Technical Summary

Problems solved by technology

[0006] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a multi-convex combined adaptive filtering method based on maximum correlation entropy, which solves the technology that the convergence speed of the combined filter and the ability to track the optimal value of the prior art decrease question

Images