机构地区: 北京科技大学计算机与通信工程学院
出 处: 《电工技术学报》 2009年第7期59-64,共6页
摘 要: 提出一种基于变论域变步长LMS自适应滤波算法,该算法的步长随偏差论域的变化而变化。理论研究和仿真结果表明,该算法在控制系统对象及其逆的建模中能同时兼顾初始收敛速度、时变系统跟踪能力及收敛精度三个性能指标,可应用于自适应逆控制系统对象及其逆的建模中,使控制系统具有良好的动静态特性。并给出了基于变论域变步长LMS算法的自适应逆控制系统全局稳定性证明。 An LMS (Least Mean Square) adaptive filtering algorithm based on variable region variable step-size is presented in this paper. The step size in the algorithm varied with the variation of the region of windage. The theory research and simulation results show that the algorithm give attention to the three performance indexes in the control and inverse system modeling, which are initial convergent speed, tracking ability of time-varying system and convergence precision. The algorithm is validated in adaptive inverse control system modeling to ensure the well dynamic and static state of the control system. The global stability of the adaptive inverse control system is proved based on the variable region variable step size LMS algorithm.