# Affordable robot avatars for rehabilitation Source: https://www.youtube.com/watch?v=u5gr6uOJ99k Recap page: https://rapidrecap.app/video/u5gr6uOJ99k Generated: 2026-03-06T09:33:28+00:00 --- ## Quick Overview The presentation advocates for the development and deployment of affordable, AI-enabled social robot avatars, like the "Lil'Flo" system, to bridge gaps in healthcare, particularly for stroke survivors and the elderly, by enabling remote, high-quality clinical interactions and therapeutic support. **Key Points:** - The COVID-19 pandemic exposed critical shortages and inefficiencies in healthcare access, leading to a 154% rise in telemedicine use. - Rehabilitation therapists surveyed identified high-quality video (87% useful) and clear screens (85% useful) as essential features for effective tele-rehabilitation. - The speaker's grandmother, a stroke survivor from Jamaica who moved to the US, experienced cognitive (46-61%), motor (65%), and psychosocial (20-33%) impairments. - The proposed solution involves an AI-enabled, affordable social robot avatar (like 'Lil'Flo') that can facilitate remote care, bridging the gap between patient needs and clinician availability. - Surveyed therapists preferred social robots over non-social telepresence devices, and users found interactive games like 'Simon Says' and 'Target Touch' engaging. - The 'Lil'Flo' system design integrates machine perception, commands, and motivation loops to enhance the patient-clinician interaction, aiming to create an embodied, high-fidelity experience. - The ultimate goal is to use this technology to support stroke patients and others needing care, regardless of their location, addressing the ongoing clinician shortage. ![Screenshot at 0:09: Speaker Michelle J. Johnson introduces her talk on developing affordable robot avatars for rehabilitation, referencing her affiliation with the University of Pennsylvania.](https://ss.rapidrecap.app/screens/u5gr6uOJ99k/00-00-09.jpg) **Context:** The presentation, delivered by Michelle J. Johnson at the AI for Good Global Summit, focuses on addressing healthcare access challenges, particularly for stroke survivors and the elderly, exacerbated by the COVID-19 pandemic. The speaker uses the personal story of her grandmother, who suffered a stroke in the 1950s, to frame the necessity of remote rehabilitation solutions. The core concept revolves around using affordable, AI-enabled social robots as telepresence avatars to deliver high-quality, engaging therapeutic interactions, compensating for existing healthcare resource gaps. ## Detailed Analysis The speaker, Michelle J. Johnson, opens by setting the context: the COVID-19 pandemic caused isolation and exposed frailties in healthcare systems, leading to a massive shift toward telemedicine (up 154%). This context motivated the development of affordable robot avatars for rehabilitation. She highlights survey data showing that therapists prioritize high-fidelity video (87% essential) and clear screens (85% essential) for tele-rehabilitation. The need is personal, stemming from her grandmother's experience with stroke in the 1950s, which resulted in significant cognitive, motor, and psychosocial impairments, and difficulties with daily living activities. The proposed solution, exemplified by the 'Lil'Flo' robot, is an AI-enabled social robot designed to bridge the access gap between clinicians and patients, both in-person and remotely. This system aims to replicate the rich, multi-directional interaction of in-person care (instruction, emotion, diagnosis, motivation). The presentation shows examples of existing robotic solutions (VGo, Nao) but argues for a more embodied, social robot approach. The team surveyed therapists, finding high utility for features like high-quality video and social games, while features like automated social interaction were less preferred. The speaker then introduces the 'Desktop-Flo' system and demonstrates the capabilities of the physical robot avatar, showing how it can follow instructions from a remote therapist (via a multi-view screen) to perform exercises like 'Simon Says' and 'Target Touch'. The goal is to create an affordable, affective platform that maintains the human clinician's presence and connection with the patient, overcoming the limitations of basic video calls. ### COVID-19 Impact on Healthcare - Nursing homes shut down and families were shut in - Telemedicine usage rose 154% - Healthcare systems were inefficient and couldn't meet needs. ### Therapist Feature Preferences - High-quality video (87% extremely useful) and clear screen (85% extremely useful) top the list - Automated assessment and social games were also rated highly useful (75% and 69% respectively). ### The Problem - Stroke and cardiovascular disease survivors face cognitive, motor, and psychosocial impairments, leading to difficulty with Activities of Daily Living (ADLs). ### The Solution - Developing an AI-enabled, affordable social robot avatar (like Lil'Flo) to support stroke patients both in-person and remotely, maintaining clinician presence. ### Demonstration - The 'Desktop-Flo' system allows a remote therapist to control the robot, which can execute complex movements in response to commands, such as playing 'Simon Says' or 'Target Touch'. ![Screenshot at 0:04: Announcement slide for the 'AI for Good Global Summit' taking place in Geneva, Switzerland, July 8-11, 2025.](https://ss.rapidrecap.app/screens/u5gr6uOJ99k/00-00-04.jpg) ![Screenshot at 0:09: Title slide for the presentation: 'Affordable Robot Avatars for Rehabilitation', featuring affiliations with UPenn Perelman School of Medicine and GRASP Lab.](https://ss.rapidrecap.app/screens/u5gr6uOJ99k/00-00-09.jpg) ![Screenshot at 0:31: Slide highlighting the problem during the pandemic: people were 'SHUT IN and ALONE' due to nursing home shutdowns and isolation.](https://ss.rapidrecap.app/screens/u5gr6uOJ99k/00-00-31.jpg) ![Screenshot at 1:09: Diagram illustrating the gap between patients and clinicians, proposing that 'Technology Can Help Bridge This Gap'.](https://ss.rapidrecap.app/screens/u5gr6uOJ99k/00-01-09.jpg)