基于虚拟检测函数下的IMM-UKF机动目标跟踪  

Based on Virtual Detection Function of IMM-UKF Maneuvering Target Tracking

作　　者：徐洋[1] 徐松涛[1] 罗文涛[1] 向建军[1] 秦占师 

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《火力与指挥控制》2015年第1期37-41,共5页Fire Control & Command Control

基　　金：陕西省电子信息系统综合集成重点实验室基金资助项目(201107Y03)

摘　　要：为了有效提高对机动目标的跟踪效果,将无迹卡尔曼滤波(UKF)引入到交互多模型(IMM)算法框架内,加强状态估计精度;引入强跟踪滤波器(STF)到UKF算法中,避免对强机动目标的过大时间延迟和跟踪性能差的缺点;提出虚拟检测函数法,在跟踪过程中自适应调整"当前"统计模型的机动参数,加大模型集与目标真实运动模式匹配概率。仿真结果验证了改进算法的有效性。To improve the tracking effect of the maneuvering targets effectively,the Unscented Kalman Filter （UKF） is introduced into the framework of the Interacting Multiple Model （IMM） algorithm,enhancing the precision of state estimation. The Strong Tracking Filter （STF） to the UKF algorithm is applied to avoid the disadvantage of long-time delay and poor performance in the tracking of the strong maneuvering. The virtual detection function is proposed to make the maneuvering parameters of CS model adaptively adjust in the tracking process,which increases the probability of matching rate with the real object model. The simulation results verify the effectiveness of the improved algorithm.

关 键 词：无迹卡尔曼滤波 交互式多模型 强跟踪滤波器 虚拟检测函数 机动参数 

分 类 号：TN955.3[电子电信—信号与信息处理]