Information fusion aspects related to GMTI convoy tracking

Abstract

Tracking of ground moving vehicles with GMTI radar is a challenging task, which calls for efficient exploitation of all information sources available. For well-separated vehicles as well as for convoy targets we focus on information fusion aspects comprising both, fusion of data from multiple dislocated sensors as well as incorporation of background information (refined models of the sensor performance, road maps, terrain screening, and simple tactical rules). Under suitably formulated modeling assumptions algorithmic solutions within the context of Gaussian sum approximations are discussed. Methods originally proposed for well-separated vehicles can be embedded into an expectation-maximization approach for dealing with collectively moving convoy targets. By this in particular, early detection of a stopping event is alleviated.