非线性自识别自校准Kalman滤波方法  

A Nonlinear Self-recognition Self-calibration Kalman Filtering Method

作　　者：傅惠民[1] 杨海峰 文歆磊 FU Huimin;YANG Haifeng;WEN Xinlei(Research Center of Small Sample Technology,Beihang University,Beijing 100191,China)

机构地区：[1]北京航空航天大学小样本技术研究中心

出　　处：《控制与信息技术》2019年第5期7-11,共5页CONTROL AND INFORMATION TECHNOLOGY

基　　金：国家重点基础研究发展计划(2012CB720000);工信部2018年智能制造综合标准化项目《基于数字仿真的机械产品可靠性测试方法标准研究与试验验证》

摘　　要：针对深空探测、制导与控制、故障诊断中状态方程为强非线性(不能线性化)并且含有未知输入(系统误差)的情况,文章提出了一种非线性自识别自校准滤波方法,并分别结合秩采样和Sigma点采样方法,详细讨论了秩采样自识别自校准Kalman滤波和Sigma点采样自识别自校准Kalman滤波。该方法能对非线性状态方程中是否含有未知输入进行自动识别,在确认存在未知输入后再对该未知输入进行估计和补偿,这样既能有效消除状态方程中的系统误差影响,又能通过状态方程与量测方程的融合减小偶然误差,从而提高滤波精度。从算例结果可以看到,与无迹Kalman滤波和自适应无迹Kalman滤波相比,采用非线性自识别自校准Kalman滤波后,估计精度至少提高64%以上。In view of the strong nonlinearity of state equations and the influence of unknown inputs(systematic errors)in deep space exploration,guidance and control,and fault diagnosis,a nonlinear sei匸recognition self-calibration filtering method was proposed.According to the rank sampling and Sigma points sampling respectively,the rank sampling self-recognition self^calibration Kalman filter and the Sigma points sampling self-recognition sei卜calibration Kalman filter were discussed in detail.Firstly,the proposed method can automatically identify unknown inputs in nonlinear state equations,and then estimate and compensate their influence when there are unknown inputs,which can not only effectively eliminate the influence of systematic errors in state equations,but also reduce the random errors through the fusion of state equations and measurement equations,improving the filtering accuracy.It can be seen from the examples that the estimation accuracy is at least 64%higher than that of the unscented Kalman filter and the adaptive unscented Kalman filter.

关 键 词：非线性滤波 秩采样 Sigma点采样 自适应滤波 故障诊断 制导与控制 深空探测 

分 类 号：V448[航空宇航科学与技术—飞行器设计] O231[理学—运筹学与控制论]