基于IMM-UKF的非线性机动目标跟踪仿真研究  

Study on Tracking Simulation of Nonlinear Maneuvering Target Based on IMM-UKF

作　　者：杨凯[1] 倪龙强[1] 张丽华[1] 姚新涛[1] 屈武斌[1] 

机构地区：[1]西北机电工程研究所,陕西咸阳712099

出　　处：《火炮发射与控制学报》2010年第3期12-16,共5页Journal of Gun Launch & Control

摘　　要：在处理非线性强机动目标跟踪问题时,扩展卡尔曼滤波方法跟踪误差大、容易引起滤波发散,同时其系统模型可能随时间变化,不能应用传统的线性滤波方法以及单一状态模型来处理该类问题。假设在整个跟踪过程中,运动由水平匀速运动、俯冲运动和爬升运动3部分组成,目标量测由斜距、方位角和高低角组成,并基于此建立了目标运动状态模型和观测模型,给出了模型转换矩阵,设计了基于IMM-UKF滤波器,并进行了仿真。结果表明,该方法可以以较高精度跟踪非线性机动目标运动状态,同时具有较高的收敛速度。When dealing with tracking problems of nonlinear maneuvering target,the extended Kalman filter（EKF）method had larger tracking error and was easy to cause filtering divergence,at the same time the system model of the method maybe change with time.Thus traditional linear filtering method and unitary state model could not be used to deal with these types of problem.In the whole process of target tracking,it was assumed that the target movement was composed of three parts of horizontal constant speed movement,diving movement,climbing movement,and the target data mesurement was composed of slant range,azimuth angle and elevation angle.Based on these,the target maneuvering state model and observation model were set up,the model transition matrix was given,and the filter based on interacting multi-mode unscented Kalman filter （IMM-UKF）was designed.The simulation of target tracking was carried out.The simulation results showed that this method can be used to track nonlinear maneuvering target with the higher accuracy and improved convergence rate.

关 键 词：自动控制技术 机动目标跟踪 卡尔曼滤波器 交互多模型 

分 类 号：TJ35[兵器科学与技术—火炮、自动武器与弹药工程]