Robotized Charging System for Electric Vehicles : Design, Implementation, and Testing of a Prototype for Automated EV Plug-In Charging

This thesis presents the development and evaluation of a robotized charging system for electric vehicles. The objective of the project is to create a mobile robotic platform capable of detecting a vehicle charge port, approaching it autonomously, and performing a simulated docking procedure. With the goal of achieving a proof of concept system, with focus on integrating perception, navigation, embedded motor control, and mechanical positioning the robotic system. The final system is implemented using a Raspberry Pi 5 running Ubuntu and ROS for high level decision making. Coordination, perception, odometry, transform handling, and navigation. Low level wheel and linear axis control are delegated to two Raspberry Pi Pico 2 W microcontrollers. Charge port detection is performed using a machine trained YOLO-based object detection model, while lidar and camera information are used during the approach and alignment stages. The two linear axis arms are used to adjust a simulated charging connector toward the detected charge port. The completed system demonstrates repeatable simulated docking, with a standard deviation of 0.77 cm, indicating that the achieved approach can produce consistent alignment. The remaining accuracy deviations are attributed to accepted alignment errors. The results show that a robot based autonomous charging approach is plausible, while further work is required to improve robustness, safety, precision, and practical suitability for real world implementation