基于随机集的多目标跟踪算法的性能评估

Performance Evaluation of Random Sets Based Multi-Target Tracking Algorithm

摘要:

给出在目标数变化情况下,利用随机集实现多目标跟踪算法性能评估方法,利用Kullback-Leibler距离度量算法来反映目标状态的信息量并进行了相应的仿真实验。实验表明,算法可稳健跟踪变目标数情况,Kullback-Leibler测度可反映算法的稳健性。

Abstract:

This paper is concerned with the performance evaluation of multi-target tracking with random sets when the number of targets varying with time. Kullback-Leibler entropy is used to measure the algorithm’s instantaneous, over-all competence in estimating the number and state of the targets, relative to complete knowledge about ground truth. Simulations show that the algorithm can track various number multi-target robustly, and Kullback-Leibler measure can reflect the robustness of the algorithm.