A low-cost lane-level navigation algorithm based on visual information.

This paper proposes a low-cost lane-level positioning algorithm. The algorithm introduces visual information on the basis of IMU/odometer/GPS fusion to correct the lateral accuracy and realize lane-level navigation. We divide visual information fusion into three cases to solve the problem of harsh scenes. To verify the effectiveness of the algorithm, we conducted tests on expressways in more than 70 cities in China, with a total test mileage of more than 50,000 kilometers. The experimental results show that the lane matching accuracy is better than 95%.