基于容积原则的概率假设密度滤波算法

为改善多目标跟踪问题中概率假设密度滤波精度与算法运行时间之间的关系,提高目标状态和数目的实时估计性能,提出了基于容积原则的概率假设密度滤波算法. 该算法在高斯混合粒子概率假设密度的框架下,利用容积数值积分原则直接计算非线性随机函数的均值和方差, 产生粒子滤波算法的重要性函数,实现高精度粒子的重构,来近似目标状态和数目的概率分布,并且在高斯混合概率假设密度滤波算法中进行采样和更新. 仿真验证了所提出算法的有效性,其Wasserstein误差距离优化了17.32%,目标数估计均值也提高了23.72%. 

Probability Hypothesis Density Filter Based on Cubature Rule and Its Application to Multi-Target Tracking

In order to balance the requirement for precision and real-time estimation performance when dealing with the problem of multi-target tracking in probability hypothesis density filter algorithm, an implementation method of PHD filter based on the cubature rule was proposed. In the framework of the probability hypothesis density for the Gaussian mixture particle, the new algorithm directly used cubature rule based numerical integration method to calculate the mean and covariance of the nonlinear random function by a set of the certain particles and their weights, thereby generating the importance density function of the particle filter algorithm to achieve high-precision particle reconstruction, and approximating to the target state and probability distribution of the target number. Finally the prediction and update distributions for the new algorithm were approached in the framework of the probability hypothesis density for the Gaussian mixture. Simulation results show the effectiveness of the proposed algorithm, that Wasserstein depresses 17.32% and mean of object number advances 23.72%.