Path Generation Using 3rd Polynomial Curve and Control Algorithm for Autonomous Parking

Abstract

This paper presents a vehicle control algorithm for autonomous parking system. The position and heading angle of subject vehicle or parking spot are acquired from the GPS information. Those determine the middle point. Paths which connect the subject vehicle and middle point are generated as a 3rd polynomial curve. It is successfully done by constraints including subject vehicle states, position and heading angle of middle point. The path links the middle point and the goal point follows the same generation process. The steering angle should not exceed the threshold value which is possible to be kept up by vehicles. The vehicle moves to a position near the parking lot. Then using a camera attached to the tail end vehicle, not only parking lanes are recognized in the backward situation but also the accurate middle point and goal point are determined. Automatic Transmission Shifting Control(ATSC) is applied for a gear shift. To generate the paths at various positions around the parking spot, it is verified by a simulation. Also the real size vehicle test is implemented and had a success rate higher than 97 percent.