红外数据辅助的GM-PHD滤波工程优化算法

GM-PHD Engineering Optimization Algorithm Based on Infrared Data Assistance

利用红外传感器优异的方位精度特点,解决随机有限集的高斯混合-概率假设密度(GM-PHD)算法在工程应用中存在的问题,提出一种在两方面对目标跟踪进行优化的方法。首先,红外传感器确定新生的目标起始信息,形成新生目标随机集初始值;其次,红外数据在目标跟踪过程中实时修正高斯权值,确保算法运行的稳定性;最后,经仿真验证文中算法目标跟踪效果得到大幅度提升。

This paper uses the excellent azimuth accuracy of infrared sensors to solve the problems of random finite set GM-PHD algorithm in engineering applications. A method is proposed to optimize the target tracking in two ways. First, the infrared sensor determines the starting information of the new target and forms the initial value of the random target of the new target. Second, infrared data corrects Gaussian weights in real-time during target tracking to ensure the stability of the algorithm. The simulation verified that the target tracking effect of this algorithm has been greatly improved.