Enhancing Smart Home Security Using Adaptive Access Control with Blockchain and Machine Learning

Abstract

Smart homes, equipped with interconnected IoT devices such as locks, cameras, and sensors, face critical security challenges due to the limitations of static access control mechanisms like Role-Based Access Control (RBAC) and Attribute-Based Access Control (ABAC), which lack adaptability to dynamic, multi-user environments and evolving threats. To address this problem, this research introduces a hybrid Blockchain–Machine Learning (ML) framework that ensures secure, adaptive, and context-aware access control for smart home ecosystems. The proposed system integrates IoT devices with ML algorithms, including Support Vector Machines (SVM) and Neural Networks, to predict user behaviours and dynamically adjust access permissions in real time, while Blockchain ensures immutable, decentralized, and tamper-proof logging of access events. The methodology employed a mixed approach, beginning with an extensive literature review to identify shortcomings in existing static models, followed by system design using smart contracts, caching strategies to reduce latency, and a user perception survey involving 25 participants to validate acceptance and usability. Results demonstrated high user trust and readiness to adopt the proposed system, with 96% of respondents favouring Blockchain-ML-enabled dynamic access control over conventional methods despite concerns about privacy risks, costs, and implementation complexity. This work contributes to society by offering a scalable and intelligent smart home security solution that enhances trust, improves user experience, and strengthens resilience against cyber threats, ultimately supporting safer and smarter living environments.