The main objective of this thesis was to develop a modular perception and control system for autonomous vehicles, as well as validate it via testing in a realistic simulation environment (Gazebo/ROS), utilizing Ackerman steering kinematics. The primary focus of the research was to achieve robust real-time autonomous navigation on complex simulated tracks through the integration of computer vision and robotics. A central focus of the research was the Hybrid Computer Vision Pipeline, which provides the ability to balance high accuracy and computational efficiency for the vehicle. For lane-following, the system utilized fast heuristic methods, specifically HSL Color Segmentation and Canny Edge Detection to provide the vehicle's geometric data (e.g., curvature). Traffic sign detection was accomplished through a multi-step process consisting of Hough Gradient Methods and Haar Cascades to quickly localize signs, then a custom CNN to accurately classify them. Overall, the research demonstrated the critical tradeoff between efficiency and reliability in achieving electric mobility automation.