新型粒子滤波算法在传递对准系统中的应用

New particle filtering algorithm and its application in transfer alignment systems

为了提高传递对准非线性系统状态估计中粒子滤波算法的估计精度,提出了一类应用中心差分滤波(CDDF)算法产生粒子建议分布的中心差分粒子滤波(CDDPF)算法.该算法应用Stirling插值公式逼近非线性函数,用Cholesky分解确保误差方差阵正定性,获得滤波稳定数值计算;应用CDDPF算法生成粒子建议分布,能够融合最新量测信息;最后应用新算法对传递对准系统模型进行最优滤波,CDDPF算法数值计算稳定性优于UKPF算法,状态变量估计精度得到明显提高.

In order to improve the estimated precision of particle filtering algorithm in transfer alignment（TA）nonlinear systems with large initial misalignment angles,the central divided difference particle filtering（CDDPF）algorithm is developed,which makes use of the central divided difference filtering（CDDF）algorithm to generate the particles set of the proposal distribution.It uses Stirling interpolation to approximate nonlinear system equations and/or measurement equations which can be easy to implement,and the Cholesky factorization of prediction error variance matrix to employe to ensure the positive definiteness of the estimation error variance matrix.These particles generated can integrate the latest measurement information into system state transition density.Finally the experiments on TA nonlinear system are conducted with the CDDPF algorithm.The simulation results indicate that,comparing to UKPF（unscented Kalman particle filtering）algorithm,the new algorithm has better numerical stability and its fstimation precision is improved obviously.