Precise localization of a vehicle within a driving lane by combining the vehicle trajectory with vision information

In this paper, the relative distance from the vehicle lane is obtained by fusing a commercialized LDWS and a low-cost GPS sensor. A virtual lane boundary is generated and used to estimate a high precision localization relative to the vehicle lane. The generated virtual lane boundary classified a straight or a curve road with lane classification algorithm. The heading angle of the vehicle is obtained by comparing the relative distance from the virtual lane boundary with that calculated from camera image of the driving lane. The virtual lane boundary with classification of the road type is used to extract the start and end points of the lane result of real road shows position correction is well done and the proposed algorithm reduces irregularity of the low-cost GPS data.