Research On Adaptive Algorithms With Minimum Error Entropy Criterion In Non-gaussian Noise Environment

The advent of the digital information technology has greatly accelerated the efficiency of the productions in all aspects of life.In industrial fields such as biomedical instruments,automatic control systems,remote sensing satellite communications,etc.,digital signal processing has been extremely widely used.The traditional adaptive filtering algorithm is widely used in scenarios such as echo cancellation,system identification and active noise control due to its simple structure and superior performance.In the face of today's increasingly severe acoustic environment,such as non-Gaussian noise and input noise environment,the adaptive filtering algorithms are unable to meet the demand.Therefore,this paper introduces bias compensation technology,zero attracting factor,normalized subband structure and the convex combination into the MEE algorithm to further improve the performance of the original MEE algorithm in different acoustic environments.Firstly,for the problem of the input noise under the non-Gaussian environment,this paper introduces the bias compensation technology into the MEE algorithm.Through the unbiased assumption and mathematical derivation,this paper obtains the bias compensation term and proposes the bias-compensated minimum error entropy algorithm to improve the original algorithm's performance under noisy input.Secondly,for the uncertain sparsity of the unknown system,this paper applies the polynomial zero attracting factor into the BCMEE algorithm and proposes the polynomial zero attracting bias-compensated minimum error entropy algorithm,so that the original algorithm can eliminate the need for the prior knowledge of the sparsity,which further improves the robustness of the BCMEE algorithm.Lastly,the correlation of the input signal and the bottleneck with the single step size affect the performance of the traditional algorithms.Based on the normalized subband adaptive filter minimum error entropy algorithm,this paper combines convex combination and polynomial zero attracting factor to propose convex normalized subband adaptive filter minimum error entropy algorithm and normalized subband adaptive filter polynomial zero attracting minimum error entropy algorithm,which have better performance in the colored input and the sparse acoustic environment.