Applied Intelligence and Computing
Fuel Efficient Self-Driven Vehicle using CNN with V2V Communication
Authors: Vignesh Krishnamurthy, Paramesh G and Shinu Abhi
Publishing Date: 01-01-2025
ISBN: 978-81-955020-9-7
Abstract
In today’s Artificial Intelligence (AI) -driven era, transportation is being revolutionized through AI and computer vision technologies, utilizing cameras, ultrasonic sensors for detecting obstacles, GPS for pinpointing locations, and ADXL345 for directional guidance. The objective is to improve vehicle safety and fuel economy by applying Machine Learning (ML) techniques, using resources like the Carla Driving Simulator’s Lane Detection data and Vehicle-to-Vehicle (V2V) communication. Raspberry Pi and Arduino are employed for computational tasks, allowing for AI-based forecasts using models such as Convolutional Neural Network (CNN), You Only Look Once (YOLO), and OpenCV. The conversion to grayscale and the application of Canny edge detection minimizes data channels, while YOLO and CNN are tasked with image segmentation. ML algorithms autonomously steer the vehicle to the most efficient and secure route, merging navigation, obstacle deterrence, collision avoidance systems, GPS, and fuel optimization. Conforming to official traffic regulations secures passenger safety and facilitates vehicle communication for maintenance alerts and troubleshooting. Vehicle performance is further refined with location-tailored enhancements.
Keywords
Computer Vision, CNN, Obstacle Avoidance, Anti-Collision, GPS, Fuel-Efficient.
Cite as
Vignesh Krishnamurthy, Paramesh G and Shinu Abhi, "Fuel Efficient Self-Driven Vehicle using CNN with V2V Communication", In: Mukesh Saraswat and Rajani Kumari (eds), Applied Intelligence and Computing, SCRS, India, 2025, pp. 223-236. https://doi.org/10.56155/978-81-955020-9-7-22