Robust Obstacle Detection in Hilly Region
Nikitha S, S. Sachin Kumar · Research Square · 2024
Abstract This Paper presents the critical challenge of robust obstacle detection in hilly terrain, where irregular terrain poses increased risk to drivers. When navigating irregular and steep landscapes, drivers face visibility issues, increasing the risk of collisions. Existing obstacle detection systems reach their limits where obstacle can be hidden. Our proposed system includes an advanced system for hilly regions by integrating sensor technologies with advanced machine learning algorithms. It uses an Arduino microcontroller with ultrasonic sensors and a servo motor for precise obstacle detection as well as an ESP32 Module equipped with camera for capturing real-time obstacles. The YOLOv2 machine learning model process these images to accurately identify obstacle under various environmental conditions. To ensure timely warnings a Bluetooth module allows communication with the smartphone application that informs the driver of road conditions and alerts him to detected obstacles or potential collision risks. The aim of the project is to significantly improve safety by improving the systems adaptability and accuracy in hairpin bends mitigating collisions and saving lives.