一种改进的次级通道辨识算法  

An improved algorithm for identification of secondary path

作　　者：高伟鹏[1] 贺国[1] 刘树勇[1] 魏国东[2] GAO Wei-peng;HE Guo;LIU Shu-yong;WEI Guo-dong(College of Power Engineering,Naval Univ. of Engineering,Wuhan 430033,China;Zhenjiang Watercraft College,Zhenjiang 212003,China)

机构地区：[1]海军工程大学动力工程学院,武汉430033 [2]镇江船艇学院,江苏镇江212003

出　　处：《海军工程大学学报》2019年第3期60-66,共7页Journal of Naval University of Engineering

基　　金：国家自然科学基金资助项目(51579242);国家自然科学青年基金资助项目(5109253)

摘　　要：为提高次级通道的辨识精度、减小辨识误差对自适应控制的影响,以横向滤波器作为估计模型,分别应用带遗忘因子的最小二乘递推算法和变步长最小均方算法来对横向滤波器的权系数进行了更新,并对两种算法的辨识精度和控制效果进行了对比。结果表明:变步长最小均方算法的性能优于带遗忘因子的最小二乘递推算法,但变步长算法仍存在收敛速度过慢、辨识残差较大的问题。为此,提出一种改进的变步长最小均方算法,并对其进行了实验验证。实验结果表明:改进之后的变步长最小均方算法的辨识精度满足控制要求,收敛速度较快。In order to improve the identification accuracy of the secondary path and reduce the impact on adaptive control,the transversal filter is used as the estimation model in this article.The least squares recursive algorithm with forgetting factor and the least mean square algorithm with variable-step are used in the update of the coefficient.The identification accuracy of two algorithms is compared through numerical simulation,and the results show that the accuracy of the least mean square algorithm with variable-step is better than that of the least squares recursion algorithm with forgetting factor.However,problems of slow convergence rate and large identification residual still exist in the former.Thus an improved least mean square algorithm with variable-step is proposed to solve the problems.The experiment proves that the accuracy of the improved algorithm comes up to the requirements of control with better effect and faster convergence rate.

关 键 词：次级通道辨识 横向滤波器 遗忘因子 改进变步长最小均方算法 

分 类 号：O328[理学—一般力学与力学基础]