针对低信噪比条件下传统粒子滤波-检测前跟踪(PF-TBD)算法粒子初始分布针对性不强而导致起始阶段检测和跟踪性能较差的问题,提出一种粒子初始化新算法.该算法首先利用目标幅度帧间的相关性,通过多帧的幅度积累,突出目标可能的位置;然后再根据这些位置初始化粒子,增强粒子分布的针对性,使粒子快速向目标真实位置聚集.理论分析和仿真结果表明,新算法提高了起始阶段目标的检测和跟踪性能.
For the problem that the lower pertinence of particle initial distribution in the traditional particles filter(PF) track before detection(TBD) algorithm results in the worse performance on detection and track at the initial phase,in the condition of low SNR,this paper proposes a new algorithm of particle initiation.This algorithm utilizes the correlation of target amplitude frames,and highlights the possible locations of targets by accumulating the multiple frame amplitudes,and strengthen the pertinence of particle distribution to allow the particles to gather at the real location of targets fast,on the basis of those particles whose positions are initialized.Theoretical analysis and simulation results show that the new algorithm can improve the performance on detection and track of targets at the initial phase.