경로 추적용 무인운송차의 개발 및 제어

Abstract

Automatic guided vehicle (AGV) is used for delivering a desired object to the position directed by user safe and exactly instead of human-being. So, AGV is used in various environmental fields such as factories, ports, hospitals, farms, and so on. However, there are still significant difficulties in measuring attitude and position of vehicle used in the flexible manufacturing system. To solve this problem, AGV must have path-following autonomously, and new control methods for AGV in practical fields is deeply needed. This thesis presents locomotion development and control results of two-wheeled AGV. AGV is under nonholonomic constraints with no slip and pure roll. Under these conditions, kinematic modeling and dynamic modeling in Cartesian coordinates are presented. Based on these modelings, two types of controllers are designed using the Lyapunov function candidate: sliding mode controller and adaptive sliding mode controller. Firstly, sliding mode controller for AGV with known parameters is designed. To design this controller, tracking errors are defined, and two sliding surfaces are chosen to guarantee that the errors converge to zero asymptotically. Control law is designed to stabilize the sliding surfaces using Lyapunov function. Secondly, adaptive sliding mode controller for AGV with unknown upper bounds of disturbance is designed. To design this controller, tracking errors are defined, and two sliding surfaces and update law are chosen to guarantee that the errors converge to zero asymptotically. Control law is designed to stabilize the sliding surfaces using Lyapunov function. The control system to implement the designed controllers is developed. The control system is developed based on PIC microcontroller. To obtain information of system state variables, the following four sensors are used: camera sensor, one potentiometer, one angular sensor and photo sensor. Two Bluetooth communications are utilized for path-following autonomously. Finally, the simulation and experimental results are shown to prove the effectiveness of the proposed controllers.