分布式自适应多传感器多目标跟踪算法  

Distributed adaptive multi-sensor multi-target tracking algorithm

作　　者：郑佳春[1] 于浩[2] 王夙歆[2] 龙延[1] 孟凡彬[2] 

机构地区：[1]集美大学信息工程学院,厦门361021 [2]天津航海仪器研究所,天津300131

出　　处：《中国惯性技术学报》2015年第4期472-476,共5页Journal of Chinese Inertial Technology

基　　金：福建省自然科学基金项目(2013J01203);厦门市科技计划项目(3502Z20130005);天津市科技兴海项目(KJXH2013-09;KJXH2014-10);天津市海洋经济创新发展区域示范项目(CXSF2014-3)

摘　　要：为了有效提高复杂环境下的船舶多目标跟踪系统的跟踪性能,提出了一种分布式自适应多传感器多目标跟踪算法。针对分布式融合结构设计了一种在线估计的自适应分配信息系数的方法,将自适应分配算法和动态权值分配算法相结合,实现在线自适应权值分配算法,以解决融合航迹误差低和不稳定性问题。对系统进行建模与分析,对提出的分布式自适应多传感器多目标跟踪算法进行了公式推导。通过仿真表明,改进的自适应算法估计精度提高了20%,同时该方法能够提高多目标跟踪系统稳健性。A distributed adaptive multi-sensor multi-target tracking algorithm was proposed to improve the tracking performance of a ship＇s multi-target tracking system in complex environments. According to the distributed fusion structure, an online adaptive estimation method of information distribution coefficients was designed which combined the adaptive allocation algorithm with the dynamic weight assignment algorithm and realized an online adaptive dynamic weight assignment algorithm to overcome the fusion track＇s large error and unstable problem. The working principle of the distributed multi-sensor and multi-target tracking system was described, the modeling and analysis of the system was presented, and the adaptive fusion algorithm was derived. Simulation results show that the estimation precision of the proposed adaptive fusion algorithm is increased by 20%. The proposed method can also improve the robustness of the multi-target tracking system.

关 键 词：分布式融合 多目标跟踪 自适应分配 权值优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]