Bift: A Blockchain-Based Federated Learning System for Connected and Autonomous Vehicles

Machine learning (ML) algorithms are essential components in autonomous driving. In most existing connected and autonomous vehicles (CAVs), a large amount of driving data collected from multiple vehicles are sent to a central server for unified training. However, data privacy and security have become crucial during the data sharing process. Federated learning for data security has arisen nowadays, it can improve the data privacy of distribute machine learning. However, the malicious attackers can still able to attack the training process. And due to the complete reliance on the central server, federated learning is very fragile. To address the above problem, we propose Bift: a fully decentralized machine learning system combined with federated learning and blockchain to provide a privacy-preserving ML process for CAVs. Bift enables distributed CAVs to train machine learning models locally using their own driving data and then to upload the local models to get a better global model. More importantly, Bift provides a consensus algorithm named PoFL to resist possible adversaries. We evaluate the performance of Bift and demonstrate that Bift is scalable and robust, and can defend against malicious attacks.