Distributed interacting multiple model H∞ filtering fusion for multiplatform maneuvering target tracking in clutter

This paper deals with the problem of tracking a single maneuvering target from multiple platforms in the cluttered environment. A new solution based on H∞H∞ filtering is presented to relax the requirement of a prior knowledge of the noise statistics in the conventional Kalman filter. The contribution of this paper is twofold. First, the distributed H∞H∞ filtering fusion formulae for single model are developed. Second, in order to carry out distributed fusion within the multiple model framework, novel equivalent platform and global models are constructed using the best fitting Gaussian approximation approach so that the developed distributed fusion formulae can be applied directly in the fusion center. The effectiveness of the proposed algorithm is demonstrated through Monte Carlo simulations involving tracking of a highly maneuvering target in the three-dimensional (3D) experiment. The algorithm performs better in a simulated uncertain noise statistics scenario than the Kalman filtering counterpart.