Smartphone-based Telerehabilitation Platform for Post-Stroke Upper-Limb Motor Recovery with Real-Time Biofeedback and AI-Driven Progress Monitoring

Project Abstract

A smartphone-based telerehabilitation platform is developed to enhance post-stroke upper-limb motor recovery by delivering immersive, home-based therapy sessions complemented with real-time biofeedback and AI-driven progress monitoring. This study integrates motion-sensing via smartphone inertial measurement units (IMUs) and external wearable sensors to quantify kinematic performance, muscle activation, and functional task completion with high fidelity. A modular therapy engine enables personalized exercise protocols, progressive intensity scaling, and remote supervision by clinicians through a secure cloud portal, ensuring adherence, safety, and timely intervention adjustments. Real-time biofeedback employs multimodal modalities—visual cues, auditory signals, and haptic feedback—to reinforce correct movement trajectories, reduce compensatory strategies, and promote neuroplastic changes. The AI-based progress monitoring subsystem analyzes longitudinal data to generate individualized recovery trajectories, detect plateaus, and forecast recovery potential using interpretable models trained on multicenter datasets, thereby guiding clinicians in optimizing therapy plans. The platform also integrates automatic outcome assessment tools for standard clinical scales (e.g., Fugl-Meyer Upper Extremity, Box and Block Test) and patient-reported outcome measures, enabling comprehensive evaluation of motor function, dexterity, and daily activity performance. A randomized, multicenter trial with a 6- to 12-week intervention period evaluates the platform’s efficacy against conventional home exercise regimens. Primary outcomes include changes in upper-limb motor impairment, functional independence, and quality of life. Secondary outcomes address user engagement, adherence rates, data reliability, and safety metrics, including adverse event monitoring. The study develops and validates a robust data fusion framework that harmonizes heterogeneous data streams from mobile sensors, wearables, and clinician inputs, ensuring privacy-preserving analytics through edge processing and encrypted communications. The user-centered design process incorporates co-creation with stroke survivors and therapists to optimize usability, accessibility, and motivation, including adaptations for bilateral involvement, cognitive load considerations, and sensory impairments. The platform’s cloud-based analytics provide clinicians with dashboards featuring trend analyses, anomaly detection, and evidence-informed recommendations for therapy progression, while automated alert systems notify care teams of any deviations from expected recovery paths. Economic evaluation explores cost-effectiveness, scalability, and potential healthcare system benefits, such as reduced hospital readmissions and shorter rehabilitation timelines. Cross-cultural validation ensures applicability across diverse populations, with multilingual interfaces and customizable therapy modules. The research addresses challenges related to data privacy, remote supervision reliability, and technology acceptance, proposing standardized protocols for remote rehabilitation trials and data governance. Overall, the project aims to demonstrate that smartphone-enabled telerehabilitation with real-time biofeedback and AI-driven monitoring can deliver equivalent or superior motor recovery outcomes compared with traditional programs, while enhancing accessibility, personalization, and resource efficiency in post-stroke upper-limb rehabilitation.

Project Overview

What This Project Is About

---

A plain-language overview of a smartphone-based system that helps people recovering from a stroke regain use of their upper limbs. It combines guided exercises, real-time feedback, and simple progress tracking to support remote rehabilitation at home or in community settings. The project investigates how mobile devices, sensors, and artificial intelligence can work together to provide convenient, effective therapy outside traditional clinics.

---

---

The Problem It Addresses

---

Many stroke survivors have limited access to in-person rehabilitation, especially after leaving hospital. This leads to slower recovery and poorer outcomes. The project targets a gap where convenient, affordable, at-home rehab tools with meaningful feedback are not widely available, aiming to improve consistency, motivation, and measurable progress for upper-limb recovery.

---

---

Objectives of the Project

---

1. Design a smartphone app that guides upper-limb exercises.
2. Incorporate real-time biofeedback to correct movement and improve safety.
3. Use lightweight AI to monitor progress and personalize exercise difficulty.
4. Enable remote data sharing with clinicians for ongoing assessment.
5. Evaluate usability, engagement, and short-term functional gains in users.

---

---

What You Will Do Step by Step

---

1. Review relevant literature on telerehabilitation and AI feedback.
2. Define exercise library and biofeedback criteria appropriate for home use.
3. Develop the mobile app prototype and integrate sensors (e.g., phone sensors or wearables).
4. Implement real-time feedback mechanisms and AI-based progress monitoring.
5. Run a small pilot study to test usability and safety.
6. Collect and analyze movement data and user-reported outcomes.
7. Refine algorithms and user interface based on feedback.
8. Prepare documentation and a demonstration of the system to stakeholders.

---

---

Expected Outcome

A functional smartphone-based rehabilitation platform with real-time guidance, a basic AI-driven progress tracker, and evidence showing improved engagement and potential upper-limb recovery for stroke survivors. The project aims to deliver a proof-of-concept tool that can be extended in future work to larger populations and longer-term outcomes.