earticle

영어

Trajectory planning is a key to achieve safe and efficient driving of automated vehicles. We propose a trajectory-planning suite that enables the generation of a safe and comfortable trajectory through receiving cooperative messages (e.g. CAM and DENM) to detect the obstacle vehicles. We also present a collision probability model to improve the accuracy of vehicles' decision-making. Our approach mainly experiences three steps. The trajectory generation module first generates a candidate set of eligible trajectories by synthetically considering the current yaw angle, velocity, and physical characteristics of the vehicle. Then the state prediction module predicts the position and velocity of the potential obstacle vehicle with respect to the transmission delay. The trajectory selection module finally chooses a best suitable trajectory to follow in terms of the current context. Moreover, we also investigate the effects of the communication performance on the trajectory planning. The extensive simulations show that the proposed trajectory-planning scheme is capable of guiding vehicles smoothly and safely across a common traffic situation against other obstacle vehicles.