中圖分類號(hào)： TN958；O212.2 文獻(xiàn)標(biāo)識(shí)碼： A DOI：10.16157/j.issn.0258-7998.211555 中文引用格式： 欒鑄徵，俞成龍，顧兵，等. 一種時(shí)變交互多模型融合目標(biāo)跟蹤方法[J].電子技術(shù)應(yīng)用，2021，47(9)：111-116. 英文引用格式： Luan Zhuzheng，Yu Chenglong，Gu Bing，et al. A time varying IMM fusion target tracking method[J]. Application of Electronic Technique，2021，47(9)：111-116.
A time varying IMM fusion target tracking method
Luan Zhuzheng，Yu Chenglong，Gu Bing，Zhao Xiantao
The 723 Institute of CSIC，Yangzhou 225101，China
Abstract： For interacting multiple model(IMM) target tracking theory, the invariable Markov transition probability matrix is used, and the residual model is used in the model probability updating, which is lack of real-time adaptability. In this paper, we propose to update the target state distribution based on the multi model filtering results, Bayesian estimation theory and multi model tracking results, update the model probability at the next moment according to the model likelihood function, and update the Markov transition probability matrix between models with the current filtering model target state distribution likelihood function. The Monte Carlo simulation is compared with the conventional IMM method, and the strong maneuvering target and disturbed static target scenes are simulated. The results show that the track error accuracy of this method is better than that of the conventional IMM method, and it can effectively track the maneuvering target.
Key words : Markov transition probability matrix；interacting multiple model(IMM)；likelihood function；Bayesian estimation
