MoGAIT — Motion & Gait AI Intelligence Tool

MoGAIT

Motion & Gait AI Intelligence Tool · v1.0 · June 2026

🎯 Purpose

MoGAIT is an AI-powered decision support system for rehabilitation robotics and gait analysis. It uses a Variational Autoencoder (VAE) + 4-layer MLP classifier to generate realistic synthetic patient cohorts, compute multi-domain risk profiles, and provide scenario-adaptive XAI (explainable AI) analysis across 20+ clinical and mass-gathering scenarios.

Developed for the AI Hackathon for People with Disabilities, King Salman Center for Disability Research, Riyadh · June 14, 2026.

👤 Principal Investigator

Prof. Dr. Utku Köse

Computer Engineering · AI & Rehabilitation Research

Süleyman Demirel University (SDU) · Isparta, Turkey
University of North Dakota · USA
VelTech University · India
Universidad Panamericana · Mexico

IEEE Senior Member ACM Professional Member

🤖 AI Architecture

VAE Decoder (GEN_NET)
4-layer MLP: 12→32→48→32→20
Generates 20 patient parameters

Risk Classifier (RISK_NET)
4-layer MLP: 10→24→24→16→4
Calibrated clinical risk scoring

XAI Method
Finite-difference gradient sensitivity
Scenario-reweighted feature attribution

Chart Engine
Chart.js 4.4.1 (biomechanical plots)
Canvas API (simulation animation)

⚠ Clinical Disclaimer

MoGAIT generates synthetic data for research and demonstration purposes only. All patient profiles are AI-generated; clinical outcomes are illustrative estimates from literature. Not validated for clinical decision-making. Always consult qualified rehabilitation clinicians.

🔑 AI Advisor — API Keys

🔒 Security

Keys stored in session memory only. Never written to disk or localStorage. Cleared on tab close.

Claude (Anthropic)

Gemini (Google)

GPT-4o (OpenAI)

⚙️ Custom Scenario

Name

Context

Environment

Primary Focus

MoGAIT

Motion & Gait AI Intelligence Tool

AI Hackathon · King Salman Center for Disability Research · June 14, 2026

🧬 Patient Analysis

🦾 Robotics & Simulation

🎯 Scenario:

Synthetic Patient Cohort

Count:

Enter count and click Generate

All Stroke SCI CP MS Amput. Elderly TBI PD

Cohort Overview

Generate a cohort to see population-level analysis. Select any patient for individual deep-dive.

Robotics Patient Cohort

Generate cohort in Patient Analysis tab

Robotics & Simulation

Population-level robot type analysis. Select a patient for individual device matching, control algorithms, and session planning.