Identifying Requirements for the Implementation of Robot-Assisted Physical Therapy in Humanoids: A user-centered design approach

Abstract

Bimanual humanoid assistive robots can be a valuable tool to improve access to physical therapy and multi-dimensional physical status monitoring for older adults living at home. However, at present, there is no implementation of end-effector robot-assisted rehabilitation in assistive social robots. Therefore, this paper illustrates the first steps of a user-centered design approach to develop such a robot for upper limb treatments and rehabilitation. Based on observation of geriatric rehabilitation and expert interviews with physical therapists, an online survey was conducted with 87 physical therapists. The first part of the questionnaire aimed to better understand the context of use, current practices, and goals of geriatric rehabilitation. Our findings suggest that integrating exercises that combine physical and cognitive skills and aim to improve specific activities of daily living (ADLs) are critical. Secondly, a KANO analysis was conducted to prioritize 20 potential features for the robot. Among these features, assist-as-needed (AAN) control for physical exercises, voice control for general settings, and the capability to perform treatments while the patient is seated or lying down were identified as essential or “must-have” features. Third, the possibility of an autonomously conducted weekly assessment of patients’ active range of motion (ROM) and weekly to monthly muscle function testing could allow the best possible monitoring of their functional status. Generally, applying a user-centered design approach allows an interdisciplinary team tasked with developing assistive robots to establish a common objective, based on which an initial prototype can be designed in the next step.