Published Paper Details:

Autonomous vehicles rely heavily on object detection systems to navigate safely, but their performance significantly degrades under low light conditions. This project explores a novel approach that combines federated learning with advanced image enhancement techniques to improve object detection in such challenging environments. By leveraging decentralized learning across multiple vehicles, the system preserves data privacy while continuously refining detection models with diverse, real-world low-light data [3][11][13]. The proposed framework enhances detection accuracy and robustness without requiring centralized data collection, thereby advancing the safety and reliability of autonomous driving at night or in poorly lit areas. This project explores the integration of Federated Learning for object detection in autonomous vehicles operating under low-light conditions. By leveraging decentralized data from multiple vehicles, the system enhances object detection accuracy without compromising data privacy. The proposed framework adapts advanced deep learning models to handle illumination challenges, improving vehicle perception and safety in night-time or poorly lit environments. The study demonstrates how FL can empower autonomous systems with robust, collaborative learning while preserving data confidentiality across diverse driving scenarios.