闪烁噪声下目标跟踪的容积粒子滤波算法

针对闪烁噪声下非线性非高斯系统的目标跟踪问题,首先建立了闪烁噪声的数学模型;然后分析了传统粒子滤波算法的优劣点,在此基础上,引入容积卡尔曼滤波算法,重新设计粒子滤波的重要性密度函数,提出用容积粒子滤波算法来跟踪目标。最后进行了仿真分析与对比。仿真结果表明,闪烁噪声条件下,容积粒子滤波算法的跟踪误差分别是传统粒子滤波算法和无迹粒子滤波算法的1/5和1/2,有更高的跟踪精度;而运行时间仅是无迹粒子滤波算法的1/2,且跟踪稳定性更好。

Target Tracking Based on Cubature Particle Filter Algorithm in Glint Noise Environment

Aimed at the target tracking problem under nonlinear non-Gaussian glint noise, the mathematical model of gilnt noise is established firstly, and then the advantages and disadvantages of traditional particle filter algorithm is analyzed, on this basis, combined with the cubature kalman filter algorithm to redesign the importance density function of particle filter, using cubature particle filtering algorithm for target tracking. Finally has carried on the analysis and comparison of the simulation, simulation results show that, under the glint noise, cubature particle filter has higher tracking precision, the tracking error is respectively1/5 and1/2 of traditional particle filter and unscented particle filter algorithm, while the running time is only1/2 of the unscented particle filter algorithm and the tracking stability is better.