A Machine Learning Attack Resilient Authentication Protocol for AI-Driven Consumer Wearable Health Monitoring

Ghaffar, Zahid ORCID: https://orcid.org/0000-0002-5546-2689, Kuo, Wen-Chung ORCID: https://orcid.org/0000-0003-0408-8663, Mahmood, Khalid ORCID: https://orcid.org/0000-0001-5046-7766, Alturki, Nazik ORCID: https://orcid.org/0000-0002-8434-7292, Saleem, Muhammad Assad ORCID: https://orcid.org/0009-0001-0308-601X and Bashir, Ali Kashif ORCID: https://orcid.org/0000-0003-2601-9327 (2025) A Machine Learning Attack Resilient Authentication Protocol for AI-Driven Consumer Wearable Health Monitoring. IEEE Transactions on Consumer Electronics. ISSN 0098-3063

Preview

Accepted Version Available under License Creative Commons Attribution. Download (626kB) | Preview

Abstract

The Internet of Medical Things (IoMT) is transforming healthcare by integrating interconnected consumer medical devices and sensors for remote patient health monitoring (RPHM). Integrating IoMT with Artificial Intelligence (AI) enables automated diagnostics and personalized healthcare while optimizing reliability and efficiency. It transforms healthcare by enabling RPHM through interconnected medical devices, wearable sensors, consumer health devices, and healthcare infrastructure. However, wireless communication among consumer wearable devices introduces significant security and privacy concerns, making them vulnerable to machine learning-based attacks, physical tampering, and impersonation threats. Although there are several authentication protocols, many do not provide robust resilience against these emerging threats. Therefore, we propose a machine learning attack resilient authentication protocol for AI-driven consumer wearable health monitoring to address these challenges. The protocol integrates an OPUF to mitigate machine learning-based attacks. We perform formal and informal security analyses, demonstrating that the proposed protocol provides mutual authentication, anonymity, and resistance to common security threats. Furthermore, the performance evaluation shows that the protocol significantly reduces communication and computation costs compared to existing protocols.