Enhanced Vehicle Detection through Multi-Sensor Fusion Utilizing YOLO-NAS and Faster R-CNN
Adi El‐Dalahmeh, Moawiah El‐Dalahmeh, Jie Li · 2024
Multi-Sensor Fusion (MSF) is pivotal in advancing autonomous driving technologies. MSF has numerous challenges, such as data synchronization, noise reduction, and managing the high-dimensional data from diverse sensors. Despite these complexities, MSF remains crucial for enhancing perception systems in autonomous vehicles. This research paper introduces an innovative MSF approach using deep neural networks that significantly heightens vehicle detection capabilities. By integrating camera, LiDAR, and RADAR data with an uncertainty estimation framework, the system showcases superior object detection accuracy. It outshines existing methodologies in both mean Average Precision and processing speed, underpinning its aptitude for real-time applications in autonomous vehicle navigation, even in complex driving environments.