Computational vision applied to mobile robotics with position control and trajectory planning: Study and application

Abstract

This paper has the primary goal of developing position control for a track driven robot on a flat surface through computer vision. Therefore, the robot is observed by a camera fixed upon the framework. Additionally, this work uses a path planning technique which is capable of driving the robot on the surface and avoiding possible obstacles, also detected by the camera. To conduct the practical tests together with the theoretical development, the robots kinematic and dynamic models are derived. The camera is a standard smartphone with the Android operating system. The robot driving system consists of two continuous rotation servomotors controlled by an Arduino microcontroller, which is also responsible for promoting the communication between the robot and computer. The driving signals are obtained through a specific control method, and the chosen path planning technique is based on A* (A Star) algorithm. All the computational steps were implemented in MATLAB software, in exception of the microcontroller source code. Finally, the results show that the simple model and controller derived are able to achieve good performance in the problem of position control.