Self-Driving Car Model Using Sensors and AI-Based Decision-Making

This study shows the development of a miniature autonomous self-driving car combining sensor fusion, embedded control, and computer vision-based detection. An Arduino Uno is utilized for motor control and real-time sensor processing, an ESP32-CAM employs lightweight neural networks for lane detection, traffic signal recognition, and stop sign identification. The vehicle uses ultrasonic sensors to detect obstacles and infrared sensors to trace lines and edges. A buck converter is used to control power to all electrical components, ensuring stability and security. The car demonstrates autonomous navigation by integrating dynamic decision-making based on multi-sensor feedback with motor actuation with a L298N driver. This project serves as a prototype to investigate ideas like perception, planning, and control that are relevant to real-world autonomous cars. Experimental results that show effective lane following, reliable obstacle avoidance, and traffic symbol detection under controlled conditions prove the system’s promise as a low-cost test platform for autonomous mobility research.