Work

Secure access to implantable medical devices: A deep learning-driven protocol using ECG signals

Year: 2025

Type: article

Abstract: The Internet of Things in medical solutions requires high levels of privacy and security. This is especially true when an implantable medical device (IMD) that monitors or controls a patient's health conditions needs reconfiguration by accessing it through wireless programmer devices. Although existing protocols for access control in implantable medical devices (IMDs) provide some level of security, they often face significant challenges in terms of accuracy and reliability. On the other hand, leveraging biometric signals is helpful in this type of IoT device. This work introduces four deep learning-based approaches for authorization between an IMD and a nearby programmer using ECG-based representations: Mono-RI and Omni-RI, which use convolutional neural networks (CNNs) for authentication, and Mono-uRI and Omni-uRI, which incorporate Siamese networks for both authentication and replay attack detection. The results show that the proposed models outperform the state-of-the-art in terms of authentication accuracy and replay attack detection. Furthermore, the results confirm that the proposed models can be applied in various scenarios, such as normal operation modes that require high security or emergency modes that prioritize rapid access, depending on the model's complexity and performance trade-offs.

Source: Array

DOI: 10.1016/j.array.2025.100650

Authors: Amirhossein Safari, Mohsen Hooshmand, Sadegh Sadeghi, Peyman Phalevani, Nasour Bagheri, Mahmoud Jabbarpour

Institutions: Institute for Advanced Studies in Basic Sciences, Sharif University of Technology, Shahid Rajaee Teacher Training University

Language: English

Topic: Advanced Authentication Protocols Security

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

SDG: Good health and well-being

Open Access status: gold

Estimated APC paid: $1,680