FPGA Based Vehicle Collision Avoidance and Accident Warning Using Sobel Operation and Manhattan Distance Metrics

A collision avoidance system in vehicles is a safety technology which is designed to prevent or reduce accidents caused due to any obstacles such as other vehicles or pedestrians. It utilizes cameras, or other sensors to monitor the vehicle's surroundings and alerting the driver to any potential danger. Advanced driver assistance systems (ADAS) and collision avoidance works hand in hand as this enhances overall vehicle safety, reduces any driver errors and also contributes to overall safety on road. This paper presents the design and implementation of a real-time collision avoidance system using Field-Programmable Gate Arrays (FPGAs). The proposed system processes the video input from a vehicle's camera and converts the video's frame to grayscale and applies Sobel edge detection on it which helps to calculate the distance between the two vehicles using Manhattan distance formula and hence effectively detects any potential collision. It utilizes FPGA's parallel processing capabilities, reconfigurability and low power to enhance the efficiency and speed of the collision detection and response in autonomous vehicles. The experiments were performed real-time on Altera Cyclone IV-E FPGA which demonstrated effectiveness of the system as the results indicated high accuracy and reliability in distance measurement and collision warning by analyzing parameters like dynamic color variation and background noise reduction for improving the safety and performance in autonomous vehicles.