This paper1, 2 revisits the problem of maneuvering target tracking and presents a new algorithm to circumvent the exponential growth of the hypotheses (mixture elements) that arises in the optimal multiple model filter. The idea of the new scheme is to replace this increasing burden at each step by a Gaussian mixture, thus maintaining a limited number of hypotheses in the filter. Numerous comparative simulations with the IMM, both in active and passive measurement cases, show that this new approach improves significantly the tracking performance in the passive case. In the active case, on the contrary, the IMM seems to remain the best complexity-performance compromise.
Maneuvering target tracking: A Gaussian mixture based IMM estimator
2012-03-01
1190718 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
A Learning Gaussian Process Approach for Maneuvering Target Tracking and Smoothing
IEEE | 2021
|