基于随机有限集理论的多传感器目标联合检测跟踪算法

针对杂波环境中的目标检测跟踪问题,提出一种基于随机有限集理论的多传感器目标联合检测跟踪算法。算法将目标状态和量测描述为随机集合,建立考虑目标出现、目标保持、目标消失等情况的目标状态随机有限集模型,以及考虑漏检和虚警的多传感器量测随机有限集模型。将目标的联合检测跟踪问题构建为目标状态集合的贝叶斯最优估计问题,并基于随机有限集理论对该贝叶斯估计算法的递推表达式进行严格理论推导。采用序贯蒙特卡罗技术实现算法的递推滤波。仿真结果验证了该算法的有效性以及算法相对于传统基于数据关联算法的性能优势。

Multisensor Joint Target Detection and Tracking Algorithm Based on Random Finite Sets

A joint detection and tracking algorithm based on Random Finite Sets (RFS) theory is proposed for target detection and tracking in presence of clutters using multiple sensors. First, target states and measurements are described as RFS variables, RFS models of target motion including target birth, target survival and target death and of multisensor measurements including miss detection and false alarm are constructed. The joint target detection and tracking problem is then modeled as a Bayesian optimal estimation to the target state RFS and the theoretically rigorous recursive formulas for the estimation are derived using RFS theory. Finally, Sequential Monte Carlo (SMC) implementation is presented to the filter recursion. Simulation results demonstrate the effectiveness of the proposed algorithm and its significant improvement in performance over traditional association-based ones.