Expectation-maximization (EM) Algorithm Based on IMM Filtering with Adaptive Noise Covariance 1)

A novel method under the interactive multiple model (IMM) filtering framework is pre- sented in this paper, in which the expectation-maximization (EM) algorithm is used to identify the process noise covariance Q online. For the existing IMM filtering theory, the matrix Q is determined by means of design experience, but Q is actually changed with the state of the maneuvering target. Meanwhile it is severely influenced by the environment around the target, i.e., it is a variable of time. Therefore, the experiential covariance Q can not represent the influence of state noise in the maneuvering process exactly. Firstly, it is assumed that the evolved state and the initial conditions of the system can be modeled by using Gaussian distribution, although the dynamic system is of a nonlinear measurement equation, and furthermore the EM algorithm based on IMM filtering with the Q identification online is proposed. Secondly, the truncated error analysis is performed. Finally, the Monte Carlo simulation results are given to show that the proposed algorithm outperforms the existing algorithms and the tracking precision for the maneuvering targets is improved efficiently.

[1]  Vesselin P. Jilkov,et al.  Survey of maneuvering target tracking: III. Measurement models , 2001 .

[2]  Arnaud Doucet,et al.  Recursive state estimation for multiple switching models with unknown transition probabilities , 2002 .

[3]  Meir Feder,et al.  Recursive estimate-maximize (EM) algorithms for time varying parameters with applications to multiple target tracking , 1995, 1995 International Conference on Acoustics, Speech, and Signal Processing.

[4]  X. R. Li,et al.  Multiple-model estimation with variable structure. IV. Design and evaluation of model-group switching algorithm , 1999 .

[5]  Y. Bar-Shalom,et al.  Tracking a maneuvering target using input estimation versus the interacting multiple model algorithm , 1989 .

[6]  X. R. Li,et al.  Online Bayesian estimation of transition probabilities for Markovian jump systems , 2004, IEEE Transactions on Signal Processing.

[7]  X. Rong Li,et al.  Survey of maneuvering target tracking: dynamic models , 2000, SPIE Defense + Commercial Sensing.

[8]  Youmin Zhang,et al.  Multiple-model estimation with variable structure. V. Likely-model set algorithm , 2000, IEEE Trans. Aerosp. Electron. Syst..

[9]  X. R. Li,et al.  Performance Prediction of the Interacting Multiple Model Algorithm , 1992, 1992 American Control Conference.

[10]  L. Bloomer,et al.  Are more models better?: the effect of the model transition matrix on the IMM filter , 2002, Proceedings of the Thirty-Fourth Southeastern Symposium on System Theory (Cat. No.02EX540).

[11]  Vesselin P. Jilkov,et al.  Overview of multiple-model methods for maneuvering target tracking , 2003, SPIE Optics + Photonics.

[12]  Hong Jeong,et al.  Application of EM algorithm to adaptive filter for multiple target tracking , 2000, 2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100).

[13]  Thia Kirubarajan,et al.  Estimation with Applications to Tracking and Navigation: Theory, Algorithms and Software , 2001 .

[14]  X. R. Li,et al.  Survey of maneuvering target tracking: II. Ballistic target models , 2001 .

[15]  X. R. Li,et al.  Survey of maneuvering target tracking. Part I. Dynamic models , 2003 .

[16]  Meir Feder,et al.  Recursive expectation-maximization (EM) algorithms for time-varying parameters with applications to multiple target tracking , 1999, IEEE Trans. Signal Process..

[17]  T. Moon.  The expectation-maximization algorithm , 1996, IEEE Signal Process. Mag..