Machine Learning based Physical Human-Robot Interaction for Walking Support of Frail People

• Published in: 2019 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) Vol. 2019; pp. 3404 - 3407
• Main Authors: Coviello, L., Cavallo, F., Limosani, R., Rovini, E., Fiorini, L.
• Format: Conference Proceeding Journal Article
• Language: English
• Published: United States IEEE 01.07.2019
• Subjects: Activities of Daily Living; Adult; Control systems; Female; Force; Frailty; Hand Strength; Humans; Legged locomotion; Machine Learning; Machine learning algorithms; Male; Robot sensing systems; Robotics; Walking
• ISSN: 1557-170X; 1558-4615
• DOI: 10.1109/EMBC.2019.8856917

In the near future robots will permeate our daily life empowering human beings in several activities of daily living. Particular, service robots could actively support indoor mobility tasks thus to enhance the independent living of citizens. They should be able to provide tailored services to citizens to achieve higher physical human-robot interaction. Too often service robots were designed without taking into account end-users functional requirements, which can change with age and geriatric syndromes. In this paper, we present a robot smart control based on machine learning strategies and adaptable to different handgrip strengths. The smart control was implemented on ASTRO robot conceived to be a companion and to support indoor mobility, among other activities. Particularly, three smart controller strategies were implemented and tested with end users from technical and user point of view. The results show promising results that underline the proposed approach was suitable for the proposed application.