Enhancing Road Safety using Computer Vision: A multi - functional ADAS Implementation

This paper describes a low-cost multifunction Advanced Driver Assistance System (ADAS) based on Computer Vision and Sensor fusion methods comprising of lane detection (based on visual processing) and automatic braking (using sensor fusion). The system was implemented, employing a Raspberry Pi 4B board and Arduino UNO R3. Lane detection is carried out via Computer vision techniques utilizing Canny Edge Detection and Hough Transform technique. The obstacle detection feature uses ultrasonic sensors to identify obstacles and initiate braking. Trailing testing scenarios were conducted on an indoor track design meant to imitate on-road conditions by maintaining inlane driving with obstacle detection/avoidance. The lane detection approach used a past-averaged steering method to ensure smooth directional control. This prototype represents a modular and extensible platform that could be built on for future enhancements and improved embedded ADAS applications.