The project focuses on developing robust road object detection for autonomous driving under extreme weather. It uses GAN-based weather artifact removal, advanced models like YOLO and Transformers, simulation in CARLA, and real-time deployment on JetAuto with sensor fusion, improving perception accuracy, efficiency, and security in adverse conditions.