VS-IMM using road map information for a ground target tracking

Tracking a ground target, based on GMTI radar measurements, is a challenging problem. This is due particularly to the environment with a high density road traffic, terrain masks... etc, and to the high manoeuverability of ground targets. In this paper, the tracking of a manoeuvring ground target on a road in a cluttered environment is studied using map information to enhance the tracking quality. Based on road segments positions, dynamic models under road constraints are built and an optimized projection technique of the estimated state is proposed in order to keep the target position and heading on the road. Using this projection approach, a multiple model tracking algorithm is derived and its performances are studied. The derived filter is an IMM (interacting multiple model) filter with a variable structure (VS-IMM) where the models set adapts sequentially according to the target position and to the road network configuration.