An adaptive road-constrained IMM estimator for ground target tracking in GSM networks

The global system for mobile communications (GSM) networks can provide position information. However, a GSM positioning system based on current specifications faces many difficulties to yield accurate position estimate for ground target tracking. The additional information of the road network is very useful not only for restricting the target position inside the road but also for providing a potential constraint on the possible target motions. In this paper, we propose an adaptive road-constrained interacting multiple model (ARC-IMM) estimator, in which the road constraint is incorporated into a variable structure IMM (VS-IMM) mechanism as a pseudomeasurement. The module set of the ARC-IMM, not only the dynamic model but also the road constraint, is updated adaptively according to the estimated target position and the road network. In particular, the selection of the dynamic models depends also on the road constraint. The simulation results verify that the proposed approach significantly improves the estimation accuracy in comparison to an IMM estimator with directional noise, a standard IMM estimator and an extended Kalman filter (EKF). And a simulation study in the case of only two measurements available demonstrates the efficiency and robustness of the proposed approach.