多传感器多机动目标跟踪的CPHD滤波算法

CPHD filter algorithm for multiple maneuvering target tracking with multi-sensors

针对单传感器在多机动目标跟踪系统中不能很好地处理目标数目变化与突发机动的问题,提出了多传感器多机动目标跟踪的概率假设密度滤波算法.以CPHD滤波算法为理论基础,同时递推概率假设密度(PHD)函数和基数分布,避免了多目标多传感器的数据关联问题.结合自适应当前统计模型,选择3个雷达作为跟踪目标的传感器,相比于单传感器降低了信息的模糊度,提高了可信度.仿真结果比较表明了多传感器CPHD滤波算法在多目标跟踪方面的性能优势.

Cardinalized probability hypothesis density（CPHD） filter algorithm with multi-sensor for multiple maneuvering target tracking was proposed.This method was based on the CPHD filter algorithm,and the probability hypothesis density（PHD） function and the cardinality distribution were propagated at the same time to avoid the data association problem in multi-sensor for multiple target tracking.With the adaptive ＂current＂ statistical model,3 radars were chosen as target tracking sensors.Compared with the single sensor,the ambiguity of information was reduced and the reliability was advanced.The advantages of CPHD filter algorithm with multi-sensor for multiple maneuvering target tracking was verified to be more accuracy by the simulation results.