SenseSafe

SenseSafe is an AI-powered, offline-first app that helps people with disabilities navigate safely using real-time detection, voice control, and SOS alerts ensuring independence, even without internet.

Comment

THE PROBLEM

Over 1 billion people worldwide live with a disability. Despite technological progress, navigating unfamiliar environments remains a daily safety crisis.

  • 1 Billion+ people live with disabilities globally
  • 98% of top websites fail accessibility standards
  • 75% face daily accessibility barriers
  • 2× higher emergency risk for disabled individuals

THE SOLUTION

SenseSafe is a production-ready Android app that combines on-device AI, emergency alerting, voice-first interaction, and multilingual support — helping people with disabilities navigate any environment safely, independently, and without internet.

CORE FEATURES

🔍 Smart Exit Scanner Real-time AI detection of exits, doors, pathways, and stairs. Sub-100ms latency. Fully offline.

🆘 One-Tap SOS System Multi-level emergency alerts transmitting GPS location, battery status, and user ability profile to emergency contacts and first responders.

🎙️ Voice Command Interface Fully hands-free operation using 5 natural language commands processed entirely on-device. Zero cloud dependency.

🌍 Multilingual Support Full interface in 6 languages with human-reviewed emergency terminology via Azure Languages.

📍 Incident Reporting GPS-tagged community reporting of accessibility barriers with real-time admin dashboard.

HOW IT WORKS

Layer 1 — Camera feed + voice input + SOS triggers Layer 2 — On-device AI: YOLOv8n + Roboflow + OpenCV (<100ms) Layer 3 — Confidence routing: high-confidence → instant output | low-confidence → Azure Computer Vision Layer 4 — Backend: FastAPI + PostgreSQL + WebSockets on Azure

IMPACT

No existing solution provides offline, AI-powered, voice-first accessibility navigation at scale. SenseSafe targets institutional deployment across colleges, hospitals, campuses, and disaster response scenarios — aligning with global accessibility legislation (ADA, European Accessibility Act, India RPWD Act).

Adwika Vishal — Solo Developer & AI Engineer

  • Complete product vision and end-to-end execution
  • AI/ML pipeline: YOLOv8n + Roboflow + OpenCV + Azure CV hybrid inference
  • Android development: Kotlin, CameraX, ML Kit, ONNX, voice commands, SOS module
  • Backend: FastAPI, PostgreSQL, WebSockets
  • Azure cloud architecture and deployment
  • Accessibility-focused design + user research with 50+ disabled individuals

Built With

  • backend:-python
  • camerax
  • fastapi
  • ml-kit
  • mobile-development:-kotlin-(android)
  • onnx
  • opencv
  • speechrecognizer
  • websocketsdatabase:
  • yolov8n
Share this project:

Updates