基于轨迹集的衍生扩展目标多伯努利滤波算法  

Spawning extended target multi-Bernoulli filtering algorithm based on sets of trajectories

作　　者：杨英壮 蔡如华[1] 吴孙勇[1,2] 刘锦新 YANG Ying-zhuang;CAI Ru-hua;WU Sun-yong;LIU Jin-xin(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》2025年第3期1024-1034,共11页Control and Decision

基　　金：国家自然科学基金项目(62263007,62161007,62071389);广西省科技厅计划项目(桂科AA19254029,桂科AA19182007);广西省自然科学基金项目(2019GXNSFBA245072).

摘　　要：针对扩展目标多伯努利滤波器(ECBMeMBer)在复杂环境下对衍生扩展目标跟踪性能严重下降以及无法提取目标轨迹的问题,提出一种基于轨迹随机有限集(Trajectory RFS)的衍生扩展目标多伯努利滤波算法(STCBMeMBer).首先,利用轨迹多伯努利RFS(Trajectory MBer-RFS)描述多扩展目标的轨迹序列,从而为扩展目标提供连续的轨迹信息;其次,提出一种多伯努利衍生模型,通过原始扩展目标的航向角与衍生扩展目标的偏转角之间的三角函数关系式建立不同衍生扩展目标的运动方程与动力学转移模型,从而实现对衍生扩展目标质心状态与外形状态的联合估计;随后,基于轨迹MBer-RFS和所提出的多伯努利衍生模型推导并提出S-TCBMeMBer滤波器,并在线性高斯条件下给出伽玛高斯逆威沙特(GGIW)混合实现.仿真结果表明,所提出算法在杂波、漏检和噪声共存的环境下能够对衍生扩展目标进行有效跟踪,并提取扩展目标完整的轨迹信息.In addressing the challenge of degraded tracking performance,particularly when dealing with spawning extended targets and the extraction of target trajectories in complex environments,this paper introduces a spawning extended target cardinality balanced multi-target multi-Bernoulli filtering algorithm based on trajectory random finite set(S-TCBMeMBer).To enhance the representation of trajectory sequences for multiple extended targets and ensure continuous trajectory information,atrajectory multi-Bernoulli random finite set(Trajectory MBer-RFS)is first employed.Following this,a multi-Bernoulli spawning model is developed,which establishes the equations of motion and dynamic transition models for different spawning extended targets by leveraging trigonometric equations linking the orientation angle of the original extended target and the deflection angle of the spawning extended target,thus enables joint estimation of the kinematic state and shape state of the spawning extended target.With the integration of the Trajectory MBer-RFS and the developed multi-Bernoulli spawning model,the S-TCBMeMBer filter is derived and presented.Additionally,implementations under linear Gaussian conditions,specifically the Gamma Gaussian inverse Wishart(GGIW)mixture,are provided.Simulation results show the effectiveness of the proposed algorithm in tracking spawning extended targets and extracting comprehensive trajectory information even in the presence of clutter,missed detection,and noise.

关 键 词：轨迹随机有限集 衍生扩展目标 多伯努利 航向角 偏转角 伽玛高斯逆威沙特 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]