Probabilistic prediction based automated driving motion planning algorithm for lane change

This paper describes design and evaluation of a motion planning algorithm of automated vehicle for lane change. Autonomous lane change is necessary for highway automated driving. In a perception part, surrounding vehicles' states are estimated and predicted. The motion of ego vehicle is also predicted and these prediction information is utilized in motion planning. In motion planning part, driving mode, which is lane keeping or lane change, target states and constraints are decided. Lane change mode decision is determined based on surrounding vehicles states and ego vehicle states. Lane change availability is decided by the safety distance that considers relative velocity and relative position of subject and surrounding vehicles. If the ego vehicle do not perform to lane change, the most proper position is selected considering the probabilistic prediction information and the safety distance. And the longitudinal control is applied to move desired merge position. A safety driving envelope is defined based on information of surrounding vehicles behaviors and is used for control constraints. In control part, the controller is designed to obtain the desired steering angle and longitudinal acceleration using a model predictive control (MPC) with constraints. The proposed automated driving algorithm has been evaluated via computer simulation studies.