Adaptive Filtering Algorithm Improvements

Filter is the most basic but important technique in electron information domain. Since the performance of the filter some time decide the capability of the production, many county have attached importance to the theory research and production exploitation of filter. Adaptive signal processing has become a new branch in this domain with the rapid development of signal processing theory and technique. One of the most active topics is adaptive algorithm.Filter can be classified into classical filter and modern filter in many class methods and adaptive filter belongs to modern filter. Adaptive filter includes liner adaptive filter and non-liner adaptive filter divided by nature. As the design of non-liner adaptive filter is more complex than liner adaptive filter, the later is often used in practice.The basic principle of adaptive filter is to adjust the weight coefficient of the filter due to some adaptive algorithm from an optimal rule. The key of filter result lies in the algorithm better or not. We study and improve the adaptive filter algorithm based on cumulates (second-order, high-order) in this paper. We provide a new method for the filter coefficient optimization by introducing PSO algorithm into filter coefficient adjustment based on high-order.There are many kinds of adaptive filter algorithm based on second-order cumulates according to different optimization rules. The familiar algorithms are Least-mean square (LMS) algorithm and Recursive Least Squares (RLS) algorithm. LMS algorithm has broadly applied in system distinguish, signal process, noise counteract, because of its simple and easy realization. After compare some kinds of typical LMS algorithm we find there is conflict between the system trace capability and reject noise.After improve on adaptive filter algorithm based on Tongue-Like Curve...