基于演化网络模型的箱粒子CPHD群目标跟踪

提出一种基于演化网络模型和区间分析的群目标势概率假设密度(cardinalized probability hypothesis density,CPHD)滤波算法。针对传统的粒子CPHD群目标跟踪算法中粒子数多、运算量大的问题,采用箱粒子实现CPHD滤波器,减少了粒子数,降低了运算量。算法通过对群目标状态采用CPHD滤波进行预测更新,并使用所获得的群信息修正群内目标的状态,进而实现对群质心的跟踪和群目标的势估计。仿真对比实验表明,所提算法在达到与传统算法相似估计性能的条件下,大幅降低了算法的运算量,同时在强杂波环境下也具有更为突出的优势。

Box particle CPHD filter for group targets tracking based onevolution network model

This paper proposes a cardinalized probability hypothesis density (CPHD) group target tracking algorithm based on the evolving network model and interval analysis. Currently the traditional particle CPHD group target tracking algorithm causes a large number of particles and a large amount of computing. The proposed algorithm uses box particle filtering to realize CPHD filter, which decreases the number of particles and computation complexity. Through predicting and updating group target state based on CPHD filtering and modifying the target state in the group by using group information, the algorithm can realize tracking for group target centroid and obtain cardinality statistics for group targets. The simulation experiments show that, the proposed algorithm greatly reduces the computational complexity while achieving similar estimation performance with the traditional algorithm and also has more prominent advantages in strong clutter environment.