光电跟踪系统中两步非线性滤波算法研究

基于非线性最小二乘准则的两步最优估计方法(Two-step optimal estimation, TSE)在第二步中需要计算雅克比矩阵的逆,且其逆的计算经常是不存在的,从而导致滤波结果发散.因此,为改进TSE算法的稳定性,在分析了TSE算法的原理的基础上,提出了改进的TSE算法,并确定了TSE算法的中间状态向量和转换矩阵的选取原则.通过非线性测量光电跟踪系统的仿真实验验证了所提出的改进的TSE算法可以保证算法的稳定性、中间状态向量和转换矩阵的选取原则的正确性,同时也证明了此算法的性能优于扩展卡尔曼滤波和U卡尔曼滤波.

Research on Two-Step Nonlinear Filter for Electro-Optical Tracking System

The Two-step optimal estimator(TSE) is an approach based on nonlinear least-squares algorithm. Jacobian matrix inversion is often inexistent in the second step of TSE,and leads to the filter result divergence so,in order to improve the stability of TSE,an improved TSE was proposed based on the analysis of theory of TSE,and confirms the selection principle of intermediate state and transformation matrix in TSE. Simulation in nonlinear measurement electro-optical tracking system not only validates that the improved TSE can ensure the algorithm's stability and the validity of the selection principle of intermediate state and transformation matrix in TSE,but also prove the performance of this algorithm outperforms EKF and UKF.