一种类箕舌线函数的变步长归一化自适应滤波算法

在综合考虑自适应滤波算法设计中收敛速度、稳态误差、计算复杂度和跟踪性能等指标的基础上，该文提出一种类箕舌线函数的变步长归一化自适应滤波算法，用类箕舌线函数代替Sigmoid函数作为步长迭代公式，引入基于相关误差的变步长调整原则，在大大增强算法稳定性的同时大幅度提升了算法的收敛速度、跟踪性能，减小了算法的计算复杂度。在Matlab平台上分析了改进的步长函数中参数begin{document}$alpha $end{document},以及的不同取值对算法的影响，并将该文算法与已有的基于Sigmoid函数和基于箕舌线函数的变步长LMS算法进行了比较，仿真结果表明，该文算法有更快的收敛速度、更好的跟踪能力以及较小的稳态误差和较强的鲁棒性。

A Kind of Versoria Function Normalized Adaptive Filtering Algorithm

On the basis of the comprehensive consideration in the design of the indexs of the adaptive filter algorithm convergence speed, steady-state error, computational complexity and tracking performance, a kind of Versoria function normalized adaptive filtering algorithm is proposed in this paper. The class Versoria function is used instead of Sigmoid function as step iterative formula, introducing variable step size based on the relevant error adjustment principle, the stability of the algorithm is enhanced greatly. At the same time, the convergence speed and tracking performance of the algorithm is promoted and the computational complexity of the algorithm is reduced. The influence of the parameterbegin{document}$alpha $end{document},anddifferent value of the step function of algorithm is analyzed on Matlab platform. Compared with the Sigmoid function variable step size LMS algorithm and variable step size LMS algorithm based on Versoria function, and the simulation results show that this algorithm has faster convergence speed, better tracking ability, smaller steady-state error and strong robustness.