一种改进的分布式弱目标检测前跟踪算法

A Modified Distributed Track-Before-Detect Algorithm for Weak Target

针对融合粒子权重的计算实现和传感器局部粒子间融合对应关系的确定问题,本文提出了一种改进的基于粒子滤波的分布式检测前跟踪(PF-DTBD)算法。算法重新推导了融合粒子权重的计算公式,得到一个近似计算式,保证计算结果正确的同时简化了计算;设计了一种基于Gibbs采样的参数估计算法解决了多传感器局部粒子间融合对应关系的确定;依据融合粒子权重与传感器局部粒子权重间的关系引入了结合序贯概率比检验和固定样本长度似然比检验的弱目标检测算法,明确实现了弱目标的检测。与单传感器处理的仿真试验对比,验证了本文算法的有效性,降低了弱目标检测时延,提高了状态估计精度。

A modified particle filter based distributed Track-Before-Detect algorithm is presented in this paper aiming at implementing the computing of fusion particle weight and solving the correspondence among sensors＇ local particles. Formula of fusion particle weight is re-deduced and an approximate expression is gained, which insures the veracity of result and reduces complexity simultaneously;A Gibbs sampler based parameter estimation algorithm is contrived to handle the correspondence among sensors ＇ local particles; SPRT-FSS likelihood ratio detect algorithm is introduced according to the relation between fusion particle weight and sensors＇ local particle weight and how to catch the weak target is definitely achieved by it. Efficiency of our algorithm is validated by comparing with simple sensor processing ,which reduces decay of detecting weak target and improves the precision of estimation.