A human-like decision intelligence for obstacle avoidance in autonomous vehicle parking

摘要

The autonomous vehicle parking problem has drawn increased attention in recent times. It can resolve the parking-related issues that involve minimizing human driving errors and saving fuel and time. In practice, parallel parking needs extra care when dynamicity is present in the environment. Many human drivers have the expertise to take quick action when obstacles appear on their driving path. An autonomous system should possess such intelligence that can mimic human-like behavior in the presence of obstacles. This work focuses on providing an intelligent autonomous parking design using a car-like mobile robot (CLMR) that efficiently parks the vehicle in dynamic environmental conditions. A novel fuzzy-based obstacle avoidance controller is proposed that integrates sensor information into the parking problem, obtained from the surrounding of a CLMR. Ultrasonic sensors’ arrangement and their grouping provide inputs for the fuzzy system. A fuzzy-based obstacle avoidance controller can execute intelligent parking like a human by avoiding static as well as moving obstacles. The proposed work is tested in different challenging environmental conditions. Simulation results demonstrate that the proposed algorithm accomplishes autonomous parallel parking reasonably well in the presence of static and moving obstacles and can be used to park the vehicle in real-time. The proposed work helps to solve the autonomous parking problem with safety, especially in dynamic environmental conditions.