SuperAssist: A User-Assistant Collaborative Environment for the supervision of medical instrument use at home

With the rise of Transmural care, patients increasingly use medical instruments at home. Maintenance and troubleshooting greatly determines the safety and accuracy of these instruments. For the supervision of these complex tasks, we developed a user-assistant collaborative environment (U-ACE). We designed three types of personal virtual assistants that vary in dialogue style and level of autonomy. According to the scenario-based design method, we did a laboratory experiment and evaluated the influence of the different assistant types on the usability of the U-ACE. Additionally, we examined if considering personal characteristics can help better gear the assistant to the users needs. Results showed that a cooperative assistant focusing on interaction and guiding the user through the required steps, was best suited and was declared the most preferred. Collaboration with an autonomous assistant that acts autonomously led to the most efficient performance. Furthermore, personal characteristics explained variance in the experienced usability. It is recommended that, while using the U-ACE, the users select their own assistant type depending on context and personal characteristics