Case Study · Social Impact · Women's Safety · Undergrad
Aarambh
In India, vulnerable spaces are everywhere for women. Aarambh — meaning 'beginning' in Hindi — is an AI-driven safety app that automatically senses danger and responds instantly, even when the user cannot. Built during undergrad at Geetanjali Institute, this is technology in service of dignity and safety.
What does it mean to feel unsafe every day?
Violence against women is not an exception in India — it is a daily reality. Women navigate vulnerable spaces constantly: poorly lit streets, public transport, late work hours, unfamiliar cities. Existing safety apps required too many manual steps in moments where seconds matter. We asked: what if the app could sense the danger itself?
An AI-driven Android app that automatically detects danger through audio, video, image, and motion data — and triggers SOS alerts instantly, even when the user cannot manually activate them.
Women's safety is a human rights issue. Technology has the power to close the gap between distress and rescue — but only if it's fast, smart, and requires almost no manual effort in a moment of crisis.
Women across India — students, working professionals, late-night commuters — especially in areas with poor network coverage or high-risk zones. Built at Geetanjali Institute of Technical Studies, Udaipur.
SOS alert delivery rate · Time from distress detection to alert · Guardian response time · Accuracy of unsafe zone identification.
Who are we truly building for?
Two users sit at the heart of Aarambh — the woman in the moment of danger, and the guardian who needs to reach her fast.
The Woman Navigating Vulnerable Spaces
Students, working professionals, and late-night commuters moving through high-risk environments where seconds matter.
- ·SOS that triggers without manual steps
- ·Live location sharing with trusted guardians
- ·Safe route suggestions in real time
- ·Works offline and in low-network areas
- ·A way to document and report incidents safely
The Guardian
Family, friends, and trusted contacts who need to know, instantly, when someone they love is in trouble.
- ·Real-time live location access
- ·Instant SOS notifications with location details
- ·Two-way communication through the app
- ·Periodic safety check-in updates
What hurts most — and why it matters
Mapped across urgency, frequency, and impact from interviews with women across India.
| Pain point | Urgency | Frequency | Impact | Chosen |
|---|---|---|---|---|
| Manual SOS steps too slow | High | High | High | ✓ |
| No offline safety option | High | High | High | ✓ |
| Unsafe zones not identified in advance | High | High | High | ✓ |
| No anonymous community reporting | Med | High | Med | — |
| Existing apps hard to use under stress | High | High | High | — |
"The biggest failure of existing safety apps is that they require the user to be calm enough to operate them. We designed Aarambh to work even when the user cannot — automatic detection, automatic alerts, automatic protection."
Six features. One mission.
Automatic SOS Alert
Triple-click power button → instantly sends GPS location to guardians and authorities via SMS and call. No unlocking, no typing needed.
AI Threat Detection
RNNs analyze real-time device data to detect unusual behavior. CNNs process images and video to identify unsafe environments automatically.
Audio Monitoring
45-second audio snippets analyzed for distress signals — shouting, keywords like 'help'. If triggered, audio sent directly to guardians and law enforcement.
Safe Route Suggestions
ML maps safe and unsafe areas, suggests safer routes in real time. Geofencing alerts users entering high-risk zones.
Offline SOS via Satellite
SOS alerts work without internet or mobile network — via satellite communication, ensuring no woman is ever truly out of reach.
Community Safety Platform
Users anonymously share experiences, categorized using NLP to surface safety concerns and build a crowd-sourced safety map.
| Solution | Pro | Con |
|---|---|---|
| Auto SOS | Works without user action | Risk of accidental triggers |
| AI Detection | Proactive not reactive | Requires training data |
| Audio Monitoring | Captures evidence automatically | Privacy needs careful design |
| Safe Routes | Prevents danger before it happens | Depends on zone data accuracy |
| Offline SOS | Works anywhere | Satellite integration complex |
| Community Platform | Crowd-sourced intelligence | Needs moderation |
"Ship Auto SOS, AI Threat Detection, Audio Monitoring, and Safe Routes as core. Community Platform and Satellite SOS as Phase 2."
What we built. What we stood for.
Aarambh shipped as a working Android prototype with a published research paper — proof that AI, built with empathy, can close the gap between distress and rescue.
- ✦AI-powered automatic SOS — triple-click power button
- ✦RNN + CNN threat detection built and tested
- ✦45-second audio monitoring with distress analysis
- ✦Real-time GPS location sharing with guardians
- ✦Safe zone mapping and geofencing
- ✦Community experience-sharing platform with NLP
- ✦Research paper published: 'Mobile Application on Women Safety'
- ✦Co-authored with 4 teammates at Geetanjali Institute
"Aarambh means beginning. It is the beginning of a world where no woman has to choose between going where she needs to go and feeling safe getting there. We built this because technology, built with empathy and purpose, can change that — one alert, one route, one woman at a time."