Realization of fuzzy logic and deep learning based position control of autonomous symmetric mobile robot

In this thesis, a three-dimensional design of a symmetrical, modular and autonomous mobile robot, which enables its movement by passing the obstacles from the starting point to the target point, is made and its prototype and control is realized. In the first stage, while the speed controls of the motors that provide the movement of the robot are carried out with PID controller, the position control is realized with fuzzy logic controller in the second stage. In the third stage of the thesis, object detection is made with the YOLO algorithm in order to detect the obstacles in the environment and it is ensured to pass the obstacle with image processing. In this process, it carried out its movement entirely with its own hardware, without the need for any other computer or sensor. In addition, an Android application is developed to send reference values to the robot and collect data during movement without the need for any additional hardware. The motion results of the mobile robot are obtained both in the simulation environment and experimentally. Firstly, in an environment free of obstacles, the target point/points are sent to robot with the Android device and the motion of the mobil robot to target is examined within the given approach distance. In order to examine the movements in all regions of the coordinate axis, different coordinate points, including one target, two targets and three targets, are chosen as references. Later, the same procedures are done for the environment with obstacle. As a result, it has been seen that the robot reaches the referenced target points within the desired approach distances without error.