The broader impact/commercial potential of this Partnerships for Innovation - Technology Translation (PFI-TT) project includes providing access to physical therapy to large sections of society who cannot access clinic-based physical therapy for economic, geographic or health mobility reasons. The new system will focus initially on patients with stroke and various orthopedic conditions. By providing accurate, automated movement guidance to patients at home, the project aims to provide more timely, personalized physical therapy. Caregivers will have the ability to track the compliance and effectiveness of physical therapy. More effective patient engagement may lead to better health outcomes. Physical therapy clinics will be enabled to expand and scale up their businesses to reach a much larger customer base with innovative hybrid business models combining in-person, remote, and virtual therapy sessions. This type of system can enable other healthcare training applications and more meaningful video visits. The broader impacts of this project also include offering training in market-driven research, technology development, commercialization and entrepreneurship to the project team as well as broadening participation of women and underrepresented minority entrepreneurs in the commercialization effort. Thus, successful development and deployment of the virtual physical therapy system will have broad societal, commercial and educational benefits.
The project creates a virtual physical therapy system by combining innovations in machine vision, artificial intelligence and mobile video technologies to ensure deep physical therapy expertise can be made available anywhere, with high accuracy, and in a personalized and timely fashion. For many medical conditions, ensuring proper exercise techniques throughout rehabilitation is very important, and constitutes a significant portion of a treatment session. Exercise quality performed by patients at home can vary. Some exercises can be harmful if performed incorrectly, causing painful setbacks. The team seeks to develop an advanced prototype of a virtual physical therapy system which can be used by patients anywhere with just a mobile device - either for live sessions with a remote therapist or for sessions with a machine-learning-based avatar which emulates a therapist with important real-time guidance. Successful implementation of the integrated solution will involve motion tracking and force estimation algorithms using mobile device cameras and machine learning algorithms for patient performance evaluation, with understanding of contextual and human factors that may guide judgment of physical therapists. The avatar and guidance rendering algorithms will minimize both human reaction and network latency in ways that can be most effective for patient training.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
