Secure and Adaptive Mutual Authentication for Smart Homes Using Blockchain and Machine Learning

Abstract

Internet of Thing is a promising technology for creating smart home systems. Devices are being added gradually in the smart home's environments, causes the significant challenges into security, scalability, compatibility, Interoperability, etc. Traditional centralized authentication methods are not able to keep the dynamic and diverse nature of these smart environment. To address these challenges, we proposed an adaptive mutual authentication scheme with Zero Knowledge Proof and machine learning integrated within blockchain based key management and storage system. The proposed approach used Elliptic Curve Cryptography technique for key generation, a consortium blockchain for storing keys and device metadata, and a hybrid encryption scheme adaptively choosing between AES-GCM and ChaCha20-Poly1305 based on IoT device capabilities. Machine learning model is integrated to predicts the optimal cryptographic parameters, and also to ensure both security and resource efficiency. The proposed mutual authentication scheme provides a secure and scalable model for smart home systems.