IMM算法实现非线性状态估计的研究与仿真  

Research and Simulation of IMM Algorithm in Nonlinear State Estimation

作　　者：陆可[1] 肖建[1] 

机构地区：[1]西南交通大学电气工程学院,四川成都610031

出　　处：《计算机仿真》2008年第5期77-80,104,共5页Computer Simulation

基　　金：国家自然科学基金(60674057);高等学校博士学科点专项科研基金(20060613003)

摘　　要：研究交互式多模型算法的非线性状态估计性能,按照工作点把非线性系统线性化为多个子模型,建立多模型自适应状态估计器。利用Monte-Carlo仿真法将其与EKF和UKF算法在不同参数下的噪声抑止能力和鲁棒性进行了比较,并分析了马尔可夫参数和模型个数对算法性能的影响。仿真结果表明该算法能达到理想的估计精度、收敛速度、稳定性和鲁棒性,能克服单一估计器由于参数变化和外部扰动所造成的估计误差过大,甚至发散的问题,能覆盖大范围的参数不确定性。The performance of nonlinear state estimation using Interacting Multiple Model Algorithm is studied. Linear sub-models are established at each equilibrium point of the nonlinear system, and the multiple adaptive filter is presented. Monte-Carlo simulations are conducted to compare it with EKF and UKF about noise restraining capacity and robustness based on different parameters. Impacts on the performance of Markov parameters and model number are analyzed. The results demonstrate that IMM algorithm has higher estimation precision and higher convergence speed, and is more stable and robust. It can solve the problems of estimation error and divergence existing in single estimator because of the changes of parameters and environments. The algorithm can cover extensive parameter uncertainty.

关 键 词：交互式多模型 非线性 状态估计 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]