In this thesis, existing solutions in the field of driving speed profile optimization of linefollowing robots were researched. The main focus was on solutions that used a PID controller or machine learning methods to control the robot. Based on this literature analysis, a system for speed profile optimization was developed using a PID controller for the control of the robot. An existing mobile robot, which was developed at the Faculty of Mechanical Engineering at the University of Ljubljana, was used in the experiments. The optimization process developed in this study begins with the mobile robot slowly driving along the path and then reconstructing it based on the measurements from the robot's sensors. After that, a computer simulation is used to find the best speed profile by changing the driving speed parameters along the path. Finally, the most successful parameter combination, which yielded the best speed profile, is tested on the real-world robot. The effectiveness of speed profile optimization was evaluated on two real paths of different difficulty.
|
