基于粒子滤波的弱目标检测前跟踪算法

针对低信噪比条件下机动目标的检测跟踪问题,提出了一种改进型的基于多模型的粒子滤波检测前跟踪算法.由于粒子退化问题,在目标信号微弱、目标发生机动或者信号幅值波动较强势,粒子滤波的TBD算法的检测概率和跟踪精度将会下降.本算法在粒子滤波的基础之上改进,即在每次循环之前加入新粒子,新粒子的分布是由平均法和前一时刻的目标估计结果进行确定.给出了粒子滤波的TBD算法推导以及数值计算过程.仿真实验表明:基于改进型粒子滤波检测前跟踪算法能够检测低信噪比的目标.

Particle Filter-Based Track-before-Detective for Weak Target

In this paper, a multiple model particle filter algorithm is presented for the moving weak target in the low SNR environment. Because of the particle's degeneracy, the detection probability and tracking accuracy of the particle filter based track-before-detective will descend in the case of the signal of the target gets weaker, the target is maneuvering or the signal amplitude fluctuation is strong. With this improved algorithm based on the particle filter, add the new particles before each cycle, and the distribution of new particles is determined by the average method and the estimation results of latest moment. Theoretical result and numerical compulation of the partical filter-based TBD algorithm is given in this paper. The simulation experiment illustrates the improved particle filter-based TBD algorithm can detect the weak target with a low SNR.