Automobile Parts Detection and Traceability Based on Blockchain and Federated Learning

Facing a complex domestic auto parts supply chain, with tracking difficulties, this paper introduces a novel solution combining blockchain and federated learning for enhanced component detection. The proposed method facilitates decentralized model training for part identification, safeguarding supplier data privacy. Federated learning constructs models from diverse manufacturers’ data without direct sharing, while blockchain incentivizes participation and replaces central server dependency. Purchasers benefit from a robust inspection tool, tracing substandard parts swiftly via blockchain. This innovation promotes manufacturer engagement, ensures data confidentiality, and tackles fragmented dataset challenges, delivering a dependable quality assessment system for automotive components and expedient fault source pinpointing.