人体快转向运动自适应卡尔曼跟踪算法

Tracking Fast Steering Motion of Human Target Based on Adaptive Kalman Filter Algorithm

穿墙雷达对人体目标进行跟踪探测面临一系列技术难点,包括墙体反射产生的虚警、射频干扰、定位误差和目标丢失等。针对上述问题,文中基于多目标数据关联和自适应的卡尔曼滤波算法,提出一种适用于穿墙雷达探测的快转向运动人体轨迹跟踪综合处理算法,重点解决快转向运动人体目标的跟踪及多运动人体目标的轨迹交叉等问题。实验结果验证了该算法具有虚警较少、轨迹平滑、计算量小、易于工程实现、定位误差可减小50%等优点。对于反向运动的人体目标,算法也可进行有效跟踪。

Through-wall radar faces with a number of technical challenges in human target detection problems, including false alarms caused by wall reflection, radio-frequency interference, target positioning error, and target loss problem. Focused on the above issues, an integrated processing algorithm adoptable for through-wall fast steering human target tracking problem has been proposed in this paper, which is based on multi-target data association and adaptive Kalman filter algorithm. The proposed algorithm specifically aims at fast steering human target tracking problem and multi-moving target cross track problem. Simulation results demonstrate the advantages of the proposed algorithm in terms of less false alarms, smooth track, less complexity, and easy engineering, 50 decrease in target positioning error. Aversive movement can be effectively tracked by the proposed method as well.