장애물 회피를 위한 전방향 주행형 차량의 제어에 관한 연구

Abstract

This dissertation proposes a tracking controller for obstacle avoidance of an omni-directional mobile robot(OMR) using a ceiling-mounted camera system method. The tracking controller is designed for the OMR to track a desired trajectory based on backstepping method using Lyapunov function. For design the tracking controller, the following tasks are implemented in this dissertation. First, OMR, obstacles and a goal point must be separated and recognized for the OMR to avoid obstacles and reach the goal point using color recognition method and ？Small Window(SW)‟ algorithm. The SW algorithm has smaller FPS(Frame Per Second) value than the general image processing method. Second, A* algorithm is proposed to generate a trajectory from the start point to the goal point for the OMR to avoid obstacles and arrive at the goal point using information of the OMR, obstacles and the goal point obtained from the proposed image processing algorithm. Using the cost function of the proposed A* algorithm, a shortest trajectory from a start point of the OMR to a goal point is generated. iv Third, modelings for the OMR to track trajectory in 2D space is presented. First, kinematic modeling of the OMR is proposed using pseudo inverse matrix and dynamic modeling considering friction of the OMR‟s wheels are presented based on Lagrange equation. From the kinematic modeling, a backstepping controller is designed for the OMR to track the planed trajectory obtained from the A* algorithm. And the dynamic controller based on dynamic modeling is designed to make the real velocity of the OMR track the desired velocity that is the output of the kinematic controller. Forth, to implement the proposed controller, an overall system is established with a ceiling-mounted camera system, OMR and obstacles. The ceiling-mounted camera system consists of ceiling-mounted camera and server computer. The OMR consists of actuator part, sensor part and control system. The actuator part consists of four mecanum wheels and four DC motors to implement omni-directional driving. The sensor part consists of compass sensor and rotary encoder. The compass sensor is used for measuring heading angle of the OMR and each rotary encoder is used for measuring each wheel‟s velocity. The control system is developed based on 8bit microprocessor PIC18F452 and PIC18F4431. The PIC18F4431 is used as master unit, and two PIC18F452s are used as slave unit. Finally, simulation results are presented to show the effectiveness of the proposed trajectory planning algorithm and tracking controller for obstacle avoidance of the OMR.