一种基于随机集的复杂环境多目标跟踪滤波算法

A Multi-target Tracking Algorithm Based on Random Sets in Complex Environment

针对非线性系统模型中复杂环境下的多目标跟踪问题，提出了一种可应用于复杂环境中的序贯蒙特卡洛概率假设密度滤波（SMC-PHD）算法。该算法首先利用有限混合模型拟合杂波强度的空间分布并估计杂波个数，使其在杂波模型未知的情况下能够稳定跟踪目标；其次将序贯蒙特卡洛方法应用到概率假设密度滤波器中，使其在解决非线性滤波问题的同时提高了目标跟踪精度。仿真实验表明了该算法在非线性复杂环境下具有良好的跟踪性能。

According to the problem of multi-target tracking performance of nonlinear model in complex environment, a novel Sequential Monte Carlo Probability Hypothesis Density （SMC-PHD） algorithm is proposed in this paper. Firstly, this algorithm can steadily track targets in complex environment by clutter intensity with finite mixture model and estimating clutter number. Secondly, solve the problem of nonlinear filter and improve the tracking accuracy by applying sampling to the PHD filter. Simulation results show that the proposed algorithm has fitting spatial distribution of the proposed algorithm can the Sequential Monte Carlo a good tracking performance in complex nonlinear environment.