基于小波变换的快速多目标多帧多空间数据关联

Fast Multi-Scan Multi-Space Data Association for Multiple Targets Using Wavelet Transform

该文提出一种基于小波变换的快速多目标多帧多空间数据关联算法。小波变换的引入把数据关联推广到多帧情况。该算法具有不依赖先验知识的特点,在复杂杂波环境下表现出较好的关联效果。算法的关联性能在多被动传感器多目标跟踪系统中进行了评估。仿真实验表明新算法在复杂杂波环境下表现出比联合概率数据关联等算法小的计算复杂度和好的关联效果。

A fast Multi-Scan Multi-Space Data Association （MSMSDA） algorithm for multiple targets is presented based on wavelet transform in this paper. With the introduction of wavelet transform, the data association is extended to multiple scan. The MSMSDA need not any prior knowledge and plays good performance in the complex clutter environment. The performance of the MSMSDA is evaluated in the multiple passive sensor multiple target tracking system in this paper. Simulations show that the proposed new algorithm has lower computational complexity and better performance of data association than the joint probabilistic data association （JPDA）, etc., under complex clutter environment.