Local path planning algorithm for intelligent vehicles in complex scene
Aiming at the problem of real-time planning in complex road environments,a new intelligent vehicle local path planning algorithm is proposed.Firstly,the vehicle motion is decoupled into independent lateral and longitudinal motions using the Frenet coordinate system.Secondly,according to the initial configuration and target configuration of the vehicle,path set of vehicle are obtained through the polynomial method.Path feasibility check is performed to obtain candidate path set.And driving risk field(DRF)theory and minimized acceleration change rate are introduced into path planning to make the planned paths more consistent with the subjective perception of drivers.Finally,the loss function is constructed and the path that minimizes the loss function is chosen as the optimal path.The results show that the method can give more reasonable driving solution in complex scene.
万方数据
维普
