동적인 환경에서 ND 방법을 이용한 자동주차 시스템

Abstract

Automatic parking assist system is a popular technology in present car concept. In this paper, a parking method which is different from path planning method is proposed. Normally, path planning method is based on exact environment information and the environment must be static. Otherwise, vehicles may suspend or stop in parking midway. It’s also difficult to structure a path in complicated environment because of obstacles surveying. In actual parking, two or more vehicles parking together are frequently, so the situation above will become serious problem. In this paper, it uses sensors to collect information, then analyzes the information to determine vehicle how to action. Using safety weight, travel distance weight and Nsteer weight which assure the vehicle’s parking path is safe, short and control least to judge the best steering angle for car parking motion. This method compares all reachable locations and then chooses the best reachable location as motion goal. So it will have lots of calculations. For reducing calculations to improve parking efficiency, I propose some methods to find invalid points that don’t need to compute.