Implementation of PID and Fuzzy Controllers Using RGB Camera in Real Robotic Vehicle and Ros2-Gazebo Platform

This work consists of the implementation of PID and Fuzzy controllers using RGB camera in real and virtual non-holonomic vehicular robots. In addition to controlling the angular and linear velocity of the robot, the camera orientation is also controlled by a proportional controller to move the servomotor and keep it aligned in the centre of the two subscribed lines. The image is computed by using the OpenCv library to segment it, perform morphological, contour and momentum operations to obtain the object's centre of gravity and guide the robot towards that centre. The virtual robot is simulated in Gazebo and its design is based on a real GoPiGo3 robot, whose URDF was configured as to reproduce the dynamics of the real robot. To control speed and direction, PID and Fuzzy controllers were applied to both the real and virtual robot, using Ros2 libraries and configured in Ubuntu software with Python. This work simplifies the way to obtain the orientation of the real robot without the need to apply geometric methods or making approximations, but rather using the odometry of the robot with ROS 2 tools. Furthermore, thanks to the use of Gazebo Platform, it was possible to simulate and then realistically implement the dynamics of the robot using PID and Fuzzy controllers so that the robot could move satisfactorily along a two-line trajectory at variable linear and angular velocity.