Blockchain-Enabled Federated Learning for Security and Privacy in Consumer Electronics Devices

Abstract

Consumer electronics devices (CEDs) are becoming increasingly interconnected and integrated into our daily lives. Thus, the demand for seamless communication, enhanced security, and reliable performance has increased. However, the widespread adoption of these devices raises significant concerns regarding security and privacy in computing, especially when collecting and processing sensitive consumer data. To address these challenges, robust security mechanisms are necessary to protect this sensitive data. In response to these challenges, Blockchain-Enabled Federated Learning for Consumer Electronics Devices (BFLCED) is proposed to make CEDs more secure and privacy-preserving. The combination of blockchain and federated learning (FL) provides a robust solution for real-world CEDs where data privacy and security are most important. The proposed BFLCED ensures devices are authenticated and communicated securely to maintain data integrity and confidentiality during model training. It generates unique identities using the Lightweight Elliptic Curve Digital Signature Algorithm (LECDSA) and digital signatures for data integrity. Parallelly, smart contracts are employed to verify device identities & data integrity automatically and enable secure communication among devices. Data privacy is maintained during model aggregation by securely aggregating updates using encryption and multi-party computation (MPC). In the end, a security analysis is conducted to evaluate the effectiveness of the proposed mechanisms in safeguarding CEDs against potential threats and vulnerabilities. Furthermore, the proposed BFLCED transforms automation and personalization by securely connecting CEDs to our daily lives.