setup of watch
Iot device

Raksha – An AI-Based IoT Smartwatch

Inspiration

The rising concerns about personal safety, especially for women, children, and the elderly, inspired us to build Raksha—an AI-powered IoT smartwatch.
Many emergency situations, like assaults, medical crises, or kidnappings, unfold too quickly for manual intervention.
Existing safety apps require user input, but in critical moments, victims may not have time to react.

Raksha solves this problem by using AI, IoT, and real-time biometrics to detect threats and send automatic emergency alerts—without needing manual activation.

What It Does

Raksha is an AI-powered smartwatch that enhances personal safety through:

Real-time stress monitoring using HRV (Heart Rate Variability) and PVR (Pulse Volume Recording).
Speech recognition to detect threat words in conversations.
AI-powered threat detection that predicts if the user is in danger.
Instant emergency alerts sent to guardians via a dedicated Android app.
Live location tracking through Firebase Cloud Messaging (FCM).

How We Built It

Hardware: MAX30100 sensor for heart rate and pulse detection.
Machine Learning: AI models trained on HRV and PVR data to predict stress levels.
Speech Recognition: Custom AI-based voice recognition system for identifying distress words.
Cloud Integration: Firebase Cloud Messaging (FCM) for real-time notifications.
Android App: Developed in React Native, displaying alerts and GPS tracking.

Challenges We Ran Into

Data Collection & Model Training: Building a reliable HRV-based stress detection model required real-world data and fine-tuning.
Speech Recognition Accuracy: Identifying threat words in noisy environments was difficult.
Real-Time Processing: Ensuring low latency in detecting stress levels and sending alerts.
Seamless Communication: Integrating hardware, AI, and cloud messaging efficiently.

Accomplishments That We're Proud Of

Developed an AI model to predict stress levels using HRV and PVR.
Integrated speech recognition with AI-driven threat detection.
Built a fully functional Android app for real-time emergency alerts.
Achieved real-time stress monitoring and alert triggering under 5 seconds.
Created a scalable system for future health monitoring applications.

What We Learned

AI + IoT integration is key to real-time safety solutions.
Speech recognition in noisy environments is challenging but possible with fine-tuned models.
Optimizing Firebase Cloud Messaging (FCM) ensures fast, reliable notifications.
Combining physiological data and AI improves accuracy in stress detection.

What's Next for Raksha?

Enhanced AI models for more accurate stress detection and threat assessment.
Customizable distress signals (e.g., voice-activated SOS, gesture-based triggers).
Advanced health tracking (oxygen levels, temperature monitoring).
Offline functionality with Bluetooth-based emergency alerts.
Partnerships with security agencies to integrate Raksha into public safety networks.