A statistical approach to map matching using road network geometry, topology and vehicular motion constraints

This paper presents a method for reliable matching of position and orientation measurements from a standard GPS receiver to a digital map. By incorporating road network topology in the matching process using a hidden Markov model, an optimum position and orientation history can be computed from a sequence of GPS measurements. Increased robustness is achieved by introducing constraints for vehicular motion in an extended Kalman filter and by reconstructing the original road network from the digital map using cubic spline interpolation. The proposed method delivers robust matching results for standard inner-city scenarios and gives a reliable estimate of the optimal position history even for severely disturbed GPS measurements.