Design and implementation of path generation algorithm for controlling autonomous driving and parking

The autonomous valet parking technology is a vehicle and IT convergence technology which drives automatically to the target parking lot without any human intervention. To realize auto-valet parking service, the movement of the vehicle is able to be controlled and spatial environments should be recognized in real-time by sensor fusion. Most fundamental technology for autonomous driving and parking service is to generate path to goal parking lot and generate control command continuously to follow the path. The path has to be generated as considering the dynamics of the vehicle and the driving path and the parking path are generated with different data structure and algorithm because the vehicle is able to follow the parking path which is short distance under 5km/hr slowly. And the smooth curve fitting is needed to follow the driving path but not to follow the parking path. In this paper, we have designed and implemented the path generation algorithms which are able to be followed parking after driving continuously as considering vehicle kinematics.