Imagine being in a serious road incident where you're disoriented, injured, or simply unable to reach your phone. That scary reality sparked a simple question for our team: could the smartphone already sitting in your pocket act as a crucial, extra layer of safety?
That's exactly why we built AERA, which stands for Accident & Emergency Response Assistant. Created for the Bharat Builds Tour - First Commit with AWS and WeMakeDevs, AERA is our attempt to turn everyday phone sensors into a thoughtful, life-saving tool.
The Problem with Potholes
You can't just tell an app to send an SOS every time it detects a sudden jolt. Phones experience wild movements all the time for harmless reasons—from hard braking and rough roads to simply dropping the device inside the car. If a safety system panics every time you hit a pothole, it isn't actually safe. The real challenge isn't just crash detection, but figuring out what to do when things are uncertain.
Detect. Verify. Protect.
We designed AERA around a three-step philosophy:
- Detect: We use the phone's accelerometer and gyroscope, paired with location data, to look for specific temporal patterns rather than relying on a single, rigid threshold.
- Verify: We don't treat our AI's prediction as absolute truth. Instead, AERA asks a human-centered question: "Possible risk detected. Are you OK?". If you're safe, you dismiss it. If you don't respond, the system knows it's time to escalate.
- Protect: AERA steps in to reduce the actions you have to take, handling the emergency workflow to notify contacts, share your location, and confirm that help has been sent.
Meet Q, Your Intelligent Sentinel
Nobody wants to read raw machine-learning outputs like "Crash Probability: 0.87" on their screen right after a stressful event. To keep the interface simple, we introduced Q, an owl-inspired visual companion. Owls naturally communicate awareness and readiness, and Q acts as a safety sentinel rather than just a decorative mascot. Q changes states—from Watching and Analyzing to Concerned, Protecting, Reassured, and Help Sent—giving you a clear visual vocabulary of what the app is doing without the confusing tech jargon.
Building the Experience First
Right now, our mobile prototype is built using React Native, Expo, TypeScript, and Expo Sensors. We focused heavily on the user experience first because a technically correct AI detection is useless if the user is too stressed to understand the interface. Moving forward, we're looking to scale our infrastructure with AWS to handle event processing, real-world sensor data, and scalable deployment.
AERA isn't here to replace emergency services, and it doesn't assume its machine-learning predictions are flawless. The core idea is much simpler: when something looks wrong, don't immediately assume—ask. And if the person cannot answer, be prepared to act.
Loading article