D. Slavov. Mobile Robot Control Using Machine Vision Smart Camera

Key Words: Mobile robot; automatic control; computer vision; machine vision; Arduino; Pixy; line tracking.

Abstract. This paper presents ways to control a mobile robot system for two different tasks: line tracking; and finding and passing through an opening in a vertical plane. The system is constructed of a threewheel platform, Arduino microcontroller and a smart camera Pixy2 for machine vision. For the line tracking task, based on the x and y coordinates at both ends of the visible line, measured by the camera, vector gradient is calculated and weight factors are adaptively determined for generating a two-component error fed to a PID controller. The feedback control successfully drives the robot over the line even in the presence of sharp turns. For the hole-passing task, the robot locates a color-illuminated opening in a vertical plane, moves frontally and passes through it. The system finds the frontal position by repeatedly measuring the opening width while maneuvering straight and diagonally towards it. Both tasks are successfully implemented despite the variable ambient lightning conditions.