Skip to content

About

PhysioTrack: AI & VR for physiotherapy IMU & pressure sensors track exercises Real-time feedback via virtual trainer Gamified rewards boost patient adherence

Resources

Stars

8 stars

Watchers

0 watching

Forks

Repository files navigation

PhysioTrack: An Intelligent AI and IoT System for Smart Physical Rehabilitation

Logo

Python TensorFlow PyTorch ONNX OpenCV MediaPipe Unity C%23 Barracuda Oculus Raspberry Pi


Abstract

Physical rehabilitation is essential for restoring mobility and improving the quality of life after knee injuries. However, patient adherence remains a major challenge due to monotony, lack of motivation, and absence of real-time corrective feedback.

PhysioTrack is an intelligent system that integrates AI, IoT sensors, and Virtual Reality (VR) to transform rehabilitation into an interactive, adaptive, and data-driven process.

The system uses IMU and pressure sensors to capture real-time motion data, which is processed by an ONNX-converted MLP model running on the edge. VR gamification provides audiovisual guidance and boosts patient motivation.

Experimental results show:

  • 96.2% accuracy in squat classification
  • 80% improvement in user satisfaction vs traditional methods

This study demonstrates the potential of AI-driven, home-based physiotherapy systems to enhance accessibility, engagement, and recovery outcomes.


Key Features

  • Real-time motion tracking via IMU + FSR sensors
  • AI-powered feedback for exercise correction
  • Virtual Reality coach in an interactive 3D environment
  • Gamified avatar customization system
  • Performance tracking and progress reports
  • Designed for both patients & physiotherapists

Methods & Materials

Hardware Components

  • MPU-6050 (Accelerometer + Gyroscope)
  • FSR-402 (Force Sensor)
  • Raspberry Pi
  • Display Screen
  • Laptop/PC (Unity + AI Development)

Software Stack

  • AI Motion Analysis (TensorFlow / PyTorch → ONNX)
  • Unity VR Environment + Oculus Integration
  • Real-time feedback system
  • Gamification mechanics

System Pipeline

Pipeline


Experimental Results

Confusion Matrix

Confusion Matrix

Feature Heatmap

Feature Heatmap

ROC Curve

Operating Characteristic


VR Scenes & Visualization

VR Scene Samples

Scene 1
Scene 2
Scene 3


Future Work

  • Personalized rehab plans
  • Advanced VR interaction
  • Full-body motion tracking
  • AI-driven smart virtual trainer

📚 Citation

If you use PhysioTrack in your research, please cite:

Mersal, Ahmed Talaat, et al. "PhysioTrack: An Intelligent AI and IoT System for Smart Physical Rehabilitation." (2025).‏

About

PhysioTrack: AI & VR for physiotherapy IMU & pressure sensors track exercises Real-time feedback via virtual trainer Gamified rewards boost patient adherence

Resources

Stars

8 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages