End-to-End Automatic Parking Based on Proximal Policy Optimization Algorithm in Carla

With the acceleration of urbanization, the number of vehicles continues to increase. Parking in small spaces such as urban parking lots and roadside areas has become a problem in people's daily lives. Therefore, automatic parking systems have gradually become a research hotspot. In addition, with the continuous development of deep reinforcement learning (DRL), more and more research is starting to explore its application in automatic parking problems. In order to design a complete parking system, we need to consider two different parking situations: vertical parking and parallel parking. This paper designs a DRL-based proximal policy optimization (PPO) algorithm to solve the automatic parking problem. Finally, through simulation verification using the open-source simulator Carla, it is experimentally demonstrated that this method can achieve parking operations in different situations while ensuring the absolute safety of the vehicles.