Self-driving cars, designed to operate without human drivers, are efficient, accident-avoidant, and poised to revolutionize commuting with advancements in machine learning, AI, and sensor technology. This study aims to develop an intelligent system for navigating a simulated city map, with a focus on autonomously maintaining lane discipline, anticipating potential scenarios, and making informed driving decisions upon the traffic signals, lights and obstacles. This research utilizes a 1/10 scale car model, e equipped with electric driveline and steering systems. The model processes image and video data from mounted cameras, employing the Hough Transform and Canny Edge Detection algorithms to enable lane-keeping function. A TensorFlow's object detection API was built to identify objects. This methodology facilitates iterative refinement of algorithms, machine learning models, and enhances cost-efficiency and safety, significantly advancing the development of full-scale self-driving vehicles.