线性随机离散系统KF中的监控策略优化研究  

Study on optimal monitoring methodology of the KF for linear random discrete-time systems

作　　者：胡崇海[1] 姜维[1] 李一军[1] 

机构地区：[1]哈尔滨工业大学管理学院,哈尔滨150001

出　　处：《控制与决策》2010年第11期1613-1618,1624,共7页Control and Decision

基　　金：高等学校博士学科点专项科研基金课题(200802131048)

摘　　要：研究线性随机离散系统卡尔曼滤波(KF)中预测误差方差P(K|K)的递归特性,在此基础上,提出基于滤波的监控策略优化问题,对问题的各种形态进行研究,并提出相应的解决方案.首先,针对监控策略优化中存在的带约束双目标优化问题,提出一种可变种群遗传算法;然后,研究从滤波处理的角度提炼出滤波过程中的监控策略优化以及各种形态下的优化方案.相应的理论分析和仿真实验对于高精滤波估计中测量方案的拟定,具有重要的实用价值与指导意义.The recursive characteristics of estimation error variance P（K｜K） in the Kalman filter（KF） for linear random discrete-time systems with single model and single survey is studied.Based on the recursive characteristics,a optimal monitoring problem is presented,and the variety of issues and corresponding solutions on this problem are studied.In this process,a series of solutions are presented which include the variable-population genetic algorithm.The problem of optimal monitoring problem and corresponding research based on KF process are studied.So the simulation result has a significant practical and guiding value in the measurement problem of filter with high-precision and high-cost.

关 键 词：卡尔曼滤波 预测误差方差 监控策略 遗传算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]