论文部分内容阅读
密集多回波环境下对弹道多目标跟踪时,融合目标运动学信息和HRRP分类信息,研究了基于HRRP分类信息辅助跟踪的最近邻算法CATNN;由于受获取实际数据的限制,为检验算法,结合目标的仿真数据,提出了多目标跟踪和识别动态交互的仿真方法;在建立的场景中与常规的JPDA、NN跟踪算法进行比较,结果表明CATNN克服了轨迹合并和误跟的现象,具有较高的多目标跟踪性能,且产生的仿真数据满足了CATNN算法验证的需求。
Under the condition of dense multi-echo, the nearest neighbor algorithm CATNN based on the HRRP classification information aided tracking is studied when it combines the target kinematics information and the HRRP classification information. In order to obtain the actual data, Target simulation data, the simulation method of multi-target tracking and dynamic interaction recognition is proposed. Compared with the conventional JPDA and NN tracking algorithms in the established scenario, the results show that CATNN overcomes the phenomenon of trajectory merging and missed tracking with higher Of the multi-target tracking performance, and the simulation data generated to meet the CATNN algorithm validation needs.