基于交互式CPHD的多传感器多机动目标跟踪  

Multi-sensor and multi-maneuver target tracking based on interactive CPHD

作　　者：蔡如华[1] 樊向婷 吴孙勇[1,2] 王力 伍雯雯 CAI Ru-hua;FAN Xiang-ting;WU Sun-yong;WANG Li;WU Wen-wen(School of Mathematics and Computing Science,Guilin University of Electronic Technology,Guilin 541004,China;Guangxi Key Laboratory of Precision Navigation Technology and Application,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学数学与计算科学学院,广西桂林541004 [2]桂林电子科技大学,广西精密导航技术与应用实验室,广西桂林541004

出　　处：《控制与决策》2022年第1期47-57,共11页Control and Decision

基　　金：国家自然科学基金项目(11661024,61861008);广西自然科学基金项目(2016GXNSFAA380073);广西研究生教育创新计划项目(2020YCXS084)。

摘　　要：针对多传感器高速多机动目标的跟踪问题,提出一种多传感器交互式贪婪势概率假设密度(MS-IMMGreedy-CPHD)滤波器.该滤波器在预测阶段,通过交互式多模(IMM)算法对势概率假设密度(CPHD)滤波中目标的状态、势分布和运动模型同时进行预测;在滤波的更新阶段,利用贪婪(greedy)量测划分机制选取多传感器量测子集和拟分区,并通过拟分区量测子集对不同模型下CPHD预测的目标状态和势分布以及模型进行交互式更新.仿真结果表明,所提出MS-IMM-Greedy-CPHD滤波能够对高机动多目标进行稳定有效的跟踪,相较于多传感器势概率假设密度(MS-CPHD)滤波,跟踪结果的OSPA误差更小且势估计更加准确.Aiming at the tracking problem of multi-sensor high-speed and multiple maneuvering targets,a multisensor interactive greedy cardinalized probability hypothesis density(MS-IMM-Greedy-CPHD)filter is proposed.In the prediction stage,the interacting multi-mode(IMM)algorithm is used to predict the state,potential distribution and motion model of the target in CPHD filtering;in the update stage of the filter,the greedy measurement partition strategy is used to select the multi-sensor measurement subsets and quasi-partition regions,and the quasi-partition measurement subset is used to predict the target state and potential distribution under different models which is updated interactively.Simulation results show that the proposed MS-IMM-Greedy-CPHD filter can track high maneuvering multi-target stably and effectively.Compared with the multi-sensor cardinalized probability hypothesis density(MS-CPHD)filter,the OSPA error of the proposed method is smaller and the cardinalized estimation is more accurate.

关 键 词：交互式多模型 机动目标 多传感器 势概率假设密度滤波 贪婪量测划分机制 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]