Intelligent Vehicle Positioning Algorithm Based on GPS and Image Fusion 2019-01-0492
Vehicle self-localization is one of the core tasks for intelligent vehicles. Basically, localization error of GPS is about 10 m, which is insufficient for intelligent vehicle positioning. Moreover, the high cost makes INS (Inertial Navigation System) cannot be used on most vehicles. This paper proposes an accurate localization method for intelligent vehicles based on GPS and image fusion from visual map. The method aims to find the pose of current position to the nearest data collection point of visual map. Firstly, coarse localization is carried out through GPS data matching, several candidate positions are selected from visual map. Furthermore, holistic feature matching is applied to compute one candidate from GPS matching results. Finally, vehicle pose is computed by matching local features and solving Perspective-Point (PnP) problem. Localization results are refined with the computed vehicle poses. Simulations are made in a 5 km- route roadway, which are in different weather conditions and different intelligent vehicles. This process is achieved by setting different simulation parameters. Simulation results show that the mean error of localization accuracy is about 12cm and the max error is 37 cm. In addition, the proposed method has good robustness to different weather conditions. This method suggests a low-cost and accurate localization solution for intelligent vehicles.