Localization for intelligent vehicle by fusing mono-camera, low-cost GPS and map data

The localization of intelligent vehicle is an important research topic in the field of intelligent transportation systems. This paper proposes a new vehicle localization method by fusing mono-camera, low-cost GPS and map data. The basic idea is: a possible position range of the vehicle is determined by fusing low-cost GPS output and the map data; Lateral spatial information for high-accuracy localization is provided frequently by vision based lane detection module; both longitudinal and lateral spatial information for high accuracy localization is provided by vision based traffic sign detection module. The proposed method is economically feasible, is reliable and does not need any construction work or change on current traffic environments. Simulation based experiments have shown the effectiveness of the proposed method in achieving high-accuracy localization result (accuracy of centimeter-level) and have also shown apparent better performance of the proposed method over the performance of former method which is similar to the proposed method.