Access Control Architecture of Assistive Robots for Physical Activity Wellbeing Data

Abstract

Multimodal Assistive Robots (MARs) present innovative opportunities in healthcare, yet they pose significant security and privacy challenges. Our research addresses these concerns using an architecture inspired by the Authentication, Authorization, and Accounting (AAA) framework. The proposed architecture is applied to an access control system using the Robot Operating System (ROS) utilising a Fitbit tracker dataset. We introduce a fine-grained, dual approach combining policy-based and attribute-based access controls to safeguard patient data confidentiality, particularly in wellbeing contexts. This paper details our novel access control mechanism within ROS, illustrating the interaction experience between the robot and the stakeholders. The paper's findings contribute to applying a user-centric, secure access framework to a real-world use case, considering organisational policies for a multimodal assistive robot in the healthcare domain.