Abstract
Research and/or Engineering Questions/Objective
This paper presents the optimal trajectory planning considering collision risk and road constraints simultaneously for automated driving in urban environment. The automated driving technology consists of the environment perception, trajectory planning, and vehicle control. In particular, the trajectory planning is a key technology for the automated vehicle being able to determine the behavior of the vehicle by itself. In this research, a real time trajectory planning algorithm is proposed for generation of optimal trajectory to avoid static and dynamic obstacles for automated driving in urban environment.
Methodology
The proposed trajectory planning algorithm is composed of three steps. The first step is to generate trajectory candidates based on road model information and vehicle state such as position, heading and speed. The trajectory candidates are designed in the curvilinear coordinates according to two maneuvering modes: lane keeping and lane change mode. The second step is to evaluate the safety, smoothness, and consistency of the trajectory candidates using collision risk assessment which judges whether the collision occurred or not between each trajectory candidate and objects. The collision risk assessment uses static grid-map and dynamic object trajectory estimation. The third step is to select the optimal trajectory which is a collision-free trajectory among the trajectory candidates. The optimal trajectory is determined using minimum cost function and it has the velocity profile information to avoid collision with static and dynamic objects. The proposed algorithm was implemented in Hyundai automated vehicle.
Results
This research presents an algorithm to generate optimal trajectory including the velocity profile to achieve the driving safety and comfort. The proposed algorithm was verified using the test vehicle developed by Hyundai Motor Company in real traffic scenario including a roundabout, an intersection, a crosswalk, etc.
Limitations of this study
In order to improve performance of trajectory planning, precise motion estimation of surround vehicles is required. Motion estimation of surround vehicles is not matured in this study. For this reason, the motion estimation of dynamic objects such as moving vehicles and humans has to be considered using road model information and object dynamics in the next study.
What does the paper offer that is new in the field including in comparison to other work by the authors?
This research proposes the optimal trajectory planning considering collision risk and road constraints simultaneously for automated driving in urban environment.
Conclusions
In this study, we achieved the reasonable result of trajectory planning for urban automated driving to improve system performance of automated driving.
Keywords : Automated Vehicle, Local Trajectory Planning, Dynamic Obstacle, Collision Detection