电脑知识与技术
電腦知識與技術
전뇌지식여기술
COMPUTER KNOWLEDGE AND TECHNOLOGY
2012年
35期
8505-8509
,共5页
自适应跟踪%机动目标跟踪%交互多模型
自適應跟蹤%機動目標跟蹤%交互多模型
자괄응근종%궤동목표근종%교호다모형
adaptive tracking%maneuvering target tracking%IMM
针对Kalman滤波存在模型参数不精确时导致的跟踪精度差、滤波发散等问题,引入一种具有Markov转移概率的结构自适应算法——IMM滤波算法,该算法支持多个模型并行工作,通过模型概率的变化自适应调整模型,可根据不同应用环境在滤波模块选择各种线性和非线性滤波算法.Monte Carlo仿真结果表明IMM滤波算法计算量适中,滤波精度较高,跟踪性能良好,显示出较好的鲁棒性和跟踪的稳定性.
針對Kalman濾波存在模型參數不精確時導緻的跟蹤精度差、濾波髮散等問題,引入一種具有Markov轉移概率的結構自適應算法——IMM濾波算法,該算法支持多箇模型併行工作,通過模型概率的變化自適應調整模型,可根據不同應用環境在濾波模塊選擇各種線性和非線性濾波算法.Monte Carlo倣真結果錶明IMM濾波算法計算量適中,濾波精度較高,跟蹤性能良好,顯示齣較好的魯棒性和跟蹤的穩定性.
침대Kalman려파존재모형삼수불정학시도치적근종정도차、려파발산등문제,인입일충구유Markov전이개솔적결구자괄응산법——IMM려파산법,해산법지지다개모형병행공작,통과모형개솔적변화자괄응조정모형,가근거불동응용배경재려파모괴선택각충선성화비선성려파산법.Monte Carlo방진결과표명IMM려파산법계산량괄중,려파정도교고,근종성능량호,현시출교호적로봉성화근종적은정성.
According to the existing Kalman filtering model parameters lead to an imprecise tracking precision of the difference, filtering divergence problem. This paper introduces a kind of Markov transition probability structure adaptive model-IMM filter?ing algorithm. This algorithm supports multiple model parallel work. It adjusts the model adaptively by the change of model prob?ability. It can choose all kinds of linear and nonlinear filtering algorithm according to different application environment in filter?ing module. The simulation of Monte Carlo results show that the IMM filtering algorithm moderates amount of calculation, get more filtering precision and good tracking performance. It shows good robustness and tracking stability.