基于萤火虫优化粒子滤波的新型机动目标跟踪算法  

Novel target tracking method based on firefly algorithm optimized particle filter

作　　者：田梦楚[1] 薄煜明[1] 吴盘龙[1] 陈志敏[1,2] 岳聪[1] 王华[2] TIAN Meng-chu BO Yu-ming WU Pan-long CHEN Zhi-min YUE Cong WANG Hua(School of Automation, Nanjing University of Science and Technology, Nanjing 210094, China Satellite Maritime Tracking ＆ Controlling Department of China, Jiangyin 214431, China)

机构地区：[1]南京理工大学自动化学院,南京210094 [2]中国卫星海上测控部,江苏江阴214431

出　　处：《控制与决策》2017年第10期1758-1766,共9页Control and Decision

基　　金：国家自然科学基金青年基金项目(61501521);国家自然科学基金面上基金项目(61473153)

摘　　要：交互式多模型粒子滤波算法需要多个模型才能对强机动目标进行跟踪,并且粒子滤波的重采样会导致粒子贫化现象,针对该问题提出一种新型机动目标跟踪方法.该方法首先将萤火虫群体的吸引和移动机制引入粒子滤波;再将改进粒子滤波引入交互式多模型中,通过智能寻优的方式提高交互式多模型的跟踪精度和稳定性.实验结果表明,相对于IMM-PF,改进方法可以用更少的时间达到同等精度,提高了机动目标跟踪的效率.An interactive multiple model particle filter algorithm needs several models to track the strong maneuvering target, and the resampling of particle filter will lead to the impoverishment phenomenon of particles, therefore, a new kind of tracking the maneuvering target is proposed. The algorithm firstly introduces attraction and moving mechanism of glowworm group into the particle filter, and then introduce the improved particle filter into the interactive multiple model to enhance its tracking precision and stability by intelligent optimization means. Experimental results show, compared with IMM-PF, the improved method can waste less time to reach the same level of precision, and to enhance the tracking performance for maneuvering target.

关 键 词：粒子滤波 萤火虫算法 机动目标跟踪 粒子贫化 交互式多模型 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]