Towards Path Planning and Environmental Recognition for Autonomous Car Parking with Multiagent Control

In bad weather environments such as rainy or snow, car parking becomes difficult. Especially, this makes an impact to some nations since like Japan and Singapore are facing the problem of aging problems. Elder finds it difficult to park their car at night which becomes worse in bad weather. In addition, generally, car parking spaces are narrow which makes the problem more complex. In this paper, we present our preliminary analysis on the performance of different vision sensors, 2D LiDAR, and 3D LiDAR on the basis of the quality point cloud data information and feature detection. Motion planning for auto-car parking depends on the extraction of environmental information. We also present our analysis on the path planning for auto-parking in narrow space on the basis of computational time, path length, and success rate. An approach which utilizes RRT* algorithms is proposed for auto-parking in different parking scenarios with a small model car in the simulation environment.