GNSS/INS-based vehicle lane-change estimation using IMM and lane-level road map

Vehicle positioning at the lane-level is becoming an enabling factor for the location-based cooperative transportation applications. With the assistance from lane-level road map data, the developing Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS) can be effectively integrated to achieve a desired positioning performance stage. In this paper, the lane-resolution vehicle positioning in lane change scenarios is concentrated to solve the uncertainty in lane occupancy and state estimation. According to a filter-based sensor fusion scheme, the Interactive Multiple Models (IMM) based strategy is adopted. By developing kinematic models that represent the lane keeping and lane changing maneuver, lane map information is coupled to restrain the state estimation process, and provides benefits to the identification of lane occupancy. Results from field tests in the practical intersection scenario illustrate the effectiveness of the proposed IMM-based solution. The fault tolerance capability is tested under the challenging GNSS availability condition, where assistance from OBD-II interface and BDS (BeiDou Navigation Satellite System) device shows great potential in application.