关键词: 高精度地图；车道推荐开发；路网模型；SD-Map；HD-Map；DWA

Abstract: In order to meet the requirement of high-level intelligent vehicles for lane recommendation, this paper has completed the mapping relationship between road and lane based on the correlation matching which is from standard definition map (SD-Map) and high-definition map (HD-Map), proposes the feasible domain space of road network, and the calculation rule of the comprehensive lane recommendation coefficient.Based on the remaining distance of road, lane speed limit and lane congestion information, a vehicle kinematics model is established, the speed space is defined by the dynamic window method (DWA) optimization algorithm, and a comprehensive cost function based on the remaining distance of the road, lane speed limit and lane congestion information is established, and the optimal lane recommendation list is obtained through iterative optimization. Finally, the optimization results are verified through the real vehicle test. Compared with the most recommended lanes based on the static map data, it can recommend lanes more effectively and flexibly based on the remaining distance from the destination, the lane speed limit and the lane congestion state, so as to ensure the optimal driving speed of the vehicle.effectively and flexibly based on the remaining distance from the destination, the lane speed limit and the lane congestion state, so as to ensure the optimal driving speed of the vehicle.

Key words: High precision map; Lane recommendation development; Road network model; SDMap; HD-Map; DWA