Autonomous Parking
Overview
Autonomous parking combines perception and localization with motion planning and vehicle control in tight spaces, often with obstacles, occlusions, and other road users. Classical methods such as Hybrid A*, model predictive control, and optimization provide structured, physically feasible trajectories, while imitation learning and reinforcement learning aim to handle complex maneuvers and interactions. Work is evaluated in simulators such as CARLA and parking datasets, as well as on real vehicles, with measures including success, collision risk, trajectory quality, and transfer to unfamiliar spaces. A central challenge is combining flexible learned behavior with reliable safety guarantees and robust performance beyond the settings used for training.
Latest papers
23 papers1 paper in the last four weeks, down 75% on the four weeks before. 0.0% of all new papers.