基于ML背景参数估计的CDKF-CPHD多目标跟踪算法

针对低信杂比环境下的多机动目标跟踪问题,提出了一种基于极大似然(ML)背景参数估计的中心差分卡尔曼-势概率假设密度滤波(BE-CDKF-CPHD)算法。算法采用ML法实时估计重尾分布模型参数,计算检测概率和虚警概率。运用极大似然-恒虚警(MLCFAR)算法对信号进行处理,提取有效量测值,将幅值似然函数与势概率假设密度滤波器(CPHD)中的目标位置似然函数相结合,通过中心差分法递归更新得到后验均值与协方差,达到对多机动目标进行跟踪的目的。仿真结果表明,在低信杂比环境中,所提算法提高了跟踪精度与目标数目估计准确度。

A CDKF-CPHD multi-target tracking algorithm based on ML background parameter estimation

Aimed at the problem of multiple maneuvering targets tracking in low signal-to-clutter ratio backgrounds,a central difference Kalman cardinalized probability hypothesis density filter based on maximum likelihood (ML) background parameter estimation (BE-CDKF-CPHD) is proposed.The ML method is used for estimating the parameters of heavy-tailed distribution,and calculating the detection probability and false alarm probability.The maximum-likelihood constant false alarm rate (ML-CFAR) is employed to process signals.In the CPHD filter,amplitude likelihood function is combined with the likelihood function of target position of the probability hypothesis density filter.The multiple maneuvering target tracking is fulfilled by estimating the mean and variance of posterior multi-target states with central difference Kalman filter.Simulation results show that the novel algorithm improves the estimate performance of target state and number.