Save Your Grandpa

Framing the Problem

What is the problem you are trying to solve? Who does it affect?

Falls and unexpected collisions are a major safety risk, especially for older adults and other vulnerable people who may not be able to call for help quickly. A delayed response after a fall can make injuries much worse. This problem affects seniors living alone, people with mobility challenges, and eventually even workers in hazardous environments. We wanted to explore whether a low-cost wearable could detect these events in real time and trigger an immediate alert.

Idea Explanation

What is your idea? How does it fix the problem?

Our idea is a wearable distress-detection system that combines obstacle awareness and fall detection. The wearable uses sensors to monitor nearby obstacles and sudden movement changes, then classifies whether something dangerous may have happened. If it detects a fall-like event, it immediately sends an alert to a remote dashboard and triggers local warning outputs. For a real product, we imagine shrinking this into a compact wearable that could escalate beyond a dashboard and contact a voice agent or emergency-response workflow automatically.

Implementation

How do all the pieces fit together? Does your frontend make requests to your backend? Where does your database fit in?

Our system is built around two ESP32 microcontrollers. The wearable ESP32 is connected to two ultrasonic sensors and one MPU6050 IMU. It reads proximity, acceleration, and orientation data, runs fall-detection logic on-device, and sends the processed state wirelessly over ESP-NOW to a second ESP32 acting as a receiver. The receiver forwards the telemetry over UART to a laptop. On the laptop, a Python bridge ingests the serial data, stores recent telemetry and event history in SQLite, and serves a browser dashboard over HTTP. The frontend dashboard reads from that backend, shows live telemetry, changes its visual state based on warnings or emergencies, and plays a local alarm sound when a fall is detected.

Challenges

What did you struggle with? How did you overcome it?

One challenge was making the system responsive enough for a live demo while still avoiding false triggers. We had to balance sensitivity with reliability, especially because normal movement can sometimes look similar to a fall. We also ran into practical systems issues, like keeping the wireless dashboard visible across devices and dealing with network instability during testing. Another challenge was merging hardware feedback like buzzers and LEDs into the existing telemetry pipeline without breaking the dashboard data flow. We overcame these issues by simplifying the architecture, separating browser-only features from Arduino-side logic, and tuning thresholds iteratively for demo conditions.

Accomplishments

What did you learn? What did you accomplish?

We built a working end-to-end prototype that combines embedded sensing, wireless communication, backend telemetry logging, and a live browser dashboard. The system can show normal telemetry, obstacle warnings, and fall detection in real time. We also added local hardware alerts and browser-based audio alarms, which made the demo feel much more real and immediate. Beyond the technical build, we learned how to connect hardware, backend, and frontend systems into one coherent product experience instead of treating them as separate pieces.

Next Steps

The next step would be turning the breadboard prototype into a small wearable form factor with cleaner hardware packaging and more robust testing. On the software side, we would improve the fall-detection model using more real-world motion data and reduce false positives. For a real product, the biggest extension would be automatic escalation: instead of only alerting a dashboard, the wearable could connect to a voice agent that checks on the user, contacts a caregiver, or escalates to emergency services when needed. We also see potential to expand the system for assisted living, home safety, and worker-safety use cases.