A sensor guided autonomous parking system for nonholonomic mobile robots

Abstract

An automated parallel parking strategy for a car-like mobile robot is presented. The study considers general cases of parallel parking for a rectangular robot within a rectangular space. The system works in three phases. In scanning phase, the parking environment is detected by ultrasonic sensors mounted on the robot and a parking position and manoeuvring path is produced if the space is sufficient. Then in the positioning phase, the robot reverses to the edge of the parking space avoiding potential collisions. Finally, in manoeuvring phase, the robot moves to the parking position in the parking space in a unified pattern, which may requires backward and forward manoeuvres depending on the dimensions of the parking space. Motion characteristics of this kind of robots are modeled, taking into account the nonholonomic constraints acting on the car-like robot. A collision-free path is planned in reference to the surroundings. The strategy has been integrated into an automated parking system, and implemented in a modified B12 mobile robot, showing capable of safe parking in tight situations.