稀疏高斯厄米特PHD机动多目标跟踪算法

针对基于概率假设密度(probability hypothesis density,PHD)的非线性机动多目标跟踪精度低、滤波发散、目标数目估计不准确等问题,提出一种基于交互式多模型的稀疏高斯厄米特PHD算法.该算法在PHD滤波器下,采用稀疏高斯厄米特方法对目标进行状态预测和量测更新,构造一种稀疏高斯厄米特PHD滤波器;然后将交互式多模型算法融入稀疏高斯厄米特PHD滤波框架中,解决了目标机动过程中运动模式不确定的问题.仿真结果表明该算法能对机动多目标进行有效的跟踪,相比交互式多模型不敏卡尔曼PHD等滤波方法具有更高的状态估计精度,且目标数目估计更准确.

Sparse Gauss-Hermite PHD Maneuvering Multi-target Tracking Algorithm

Considering the low accuracy, filter divergence, incorrect estimation of number, and other problems of nonlinear multi-target tracking based on probability hypothesis density (PHD), a sparse Gauss-Hermite PHD algorithm based on interactive multiple models is proposed. In the proposed algorithm, a sparse Gauss-Hermite integration method is adopted for prediction and measurement update, and a sparse Gauss-Hermite PHD filter is constructed. On this basis, the motion pattern uncertainty in the target maneuvering system is solved by integrating the interactive multi-model algorithm into the sparse Gauss-Hermite PHD filtering framework. The simulation results show that the proposed algorithm has a high precision, and it is accurate in estimating the number of targets.