基于自适应GMM杂波估计的改进MHT算法

在传统多假设跟踪(MHT)算法中通常会假设杂波强度先验已知，当观测场景中杂波未知且空变时，该假设将会导致跟踪算法性能急剧下降。针对这一问题，本文提出一种基于自适应高斯混合模型(GMM)在线估计未知杂波的改进MHT算法。首先利用自适应GMM拟合未知杂波空间分布，并自适应地估计出波门内的杂波强度；然后将其应用于MHT处理中，有效改善航迹得分计算和最优假设航迹估计的准确性，进而实现在杂波未知场景中的稳定跟踪。仿真结果表明，在未知杂波观测场景中，所提算法相比传统MHT算法和MHT-GMM算法获得了更好的数据关联准确性和航迹维持性能。

An improved MHT method with clutter estimation based on adaptive Gaussian Mixture Model

In the traditional Multiple Hypothesis Tracker(MHT) algorithm, it is usually assumed that the clutter intensity is known a priori. When the clutter of observation scene is unknown and spatially variable, the performance of the tracking algorithm drops sharply. To solve this problem, an improved MHT method with clutter estimation based on adaptive Gaussian Mixture Model(GMM) is proposed. Firstly, the adaptive GMM is utilized to fit the spatial distribution of unknown clutter, and the clutter intensity in the gate is estimated adaptively. Then, it is applied to the MHT tracker to effectively improve the accuracy of track score calculation and optimal hypothetical track estimation, so as to realize stable tracking in unknown clutter scene. Simulation results show that the proposed algorithm achieves better data association accuracy and track maintenance performance than the standard MHT algorithm and the MHT-GMM algorithm in unknown clutter observation scene.