Planning flexible and smooth paths for lane‐changing manoeuvres of autonomous vehicles

Abstract

Lane-changing manoeuvres are essential for daily driving tasks. Safety and comfort are the primary issues to be considered during the lane-changing manoeuvres of autonomous vehicles. However, because the traffic environment is complicated and changeable, executing fixed lane-changing manoeuvres cannot guarantee safety and comfort. Thus, it is important to find a flexible method that can plan and re-plan a lane-changing path according to the traffic environment in real time. But that remains smooth to ensure the vehicle occupants stay comfortable. Therefore, this paper proposes a method for planning a flexible and smooth lane-changing path in real time. The proposed method uses the bi-elementary paths to ensure comfort, and clothoid and arc curves are superimposed on the bi-elementary paths to add re-planning flexibility. Additionally, lookup tables and reasonable approximations are employed to quickly generate and evaluate lane-changing paths. The results of the simulations demonstrate that the proposed method is able to plan and re-plan a smooth lane-changing path. Then, the results of vehicle experiments performed based on the embedded controller platform show that the proposed method guarantees real-time performance, that the planned path is easy to follow, and that the planned paths make autonomous vehicle lane-changing manoeuvres safe and comfortable.