Recursive parameter estimation methods based on gradient search for estimating system parameters of first-order inertial control systems
作者单位：School of Internet of Things TechnologyWuxi Vocational Institute of Commerce
会议名称：《第40届中国控制会议》
会议日期：2021年
学科分类：0711[理学-系统科学] 07[理学] 08[工学] 0835[工学-软件工程] 0802[工学-机械工程] 071102[理学-系统分析与集成] 080201[工学-机械制造及其自动化]
基 金：supported by Qing Lan Project by the”333” project of Jiangsu Province (No. BRA2018328) by Jiangsu overseas visiting scholar program for university prominent young and middle-aged teachers and presidents
关 键 词：Recursive estimation Least mean square Stochastic gradient Multi-innovation
摘 要：This paper studies the recursive parameter estimation methods for the first-order inertial system based on the impulse responses and the gradient search. By constructing a cost function with respect to system parameters in accordance with the output response and observed data, the least mean squares method, the stochastic method and the multi-innovation method are proposed. The stochastic gradient method is presented to avoid solving complicated one-dimensional equation for determining step-sizes and the multi-innovation stochastic is proposed to enhance the estimation accuracy. The simulation results are provided and show that the proposed methods are effective.
