远距离纯方位跟踪的改进粒子滤波算法

摘要

The particle filtering is a filtering method based on Monte Carlo sampling and recursive Bayesian estimation. It has been applied to various nonlinear filtering and parameter estimation, for it is not constrained by nonlinearity and non-Gaussian noise of system. As for the problems of few priori information or low SNR, the huge calculation and slow efficiency limited its application in real-time systems. The basic theory of particle filtering is introduced. The calculation speed was improved by modifying the weights calculation formula and re-sampling algorithm. Numerical simulations demonstrate that the modified method can improve the tracking performance and tracking rate for far-distance bearings-only targets.%粒子滤波是一种基于蒙特卡罗采样和递推贝叶斯估计的滤波方法,不受模型非线性和非高斯噪声的限制,因而被用于各种非线性滤波和参数估计问题.但是对于先验信息较少和信噪比较低的问题,其庞大的计算量和缓慢的速度限制了它在实时系统中的应用.这里介绍了粒子滤波基本原理,通过改进权重计算、重采样算法,使计算速度得到提高.这种改进算法用于对海洋远距离纯方位目标进行跟踪仿真,其结果表明,具有速度快,精度高的特点.