FIeld-Hive

Inspiration

Field-Hive was inspired by emergency response situations where responders need fast, simple, and portable health signals. We wanted a system that can monitor vital signs and movement in real time, then share that data wirelessly to support faster rescue decisions.

What it does

Field-Hive collects and transmits key wellness signals from a person in the field:

Heart activity from a pulse sensor (BPM estimation)
Activity/motion from an MPU6050 accelerometer
Camera-based heart and breathing metrics through a vitals API pipeline

The system sends data through ESP-NOW and UDP so a remote station can monitor status and detect potential risk conditions.

How we built it

We built Field-Hive as a multi-part embedded + networking stack:

ESP32 firmware for pulse sampling, filtering, beat detection, and BPM output
ESP32 firmware for accelerometer capture and movement classification
ESP-NOW communication between sensor nodes for low-latency local telemetry
OLED display integration for on-device status feedback
A C++ camera-vitals client (hello_vitals.cpp) using a SmartSpectra/Presage API flow
UDP JSON messaging for forwarding vitals to a receiver endpoint

Challenges we ran into

Signal noise and instability in pulse readings, especially during movement
Avoiding false beat detection and handling periods with no reliable pulse
Synchronizing different data rates (fast sensor sampling vs. slower display/transmit cycles)
Setting up reliable wireless communication and peer MAC/key configuration
Maintaining stream resilience and reconnect behavior for camera-based vitals

Accomplishments that we're proud of

Built an end-to-end prototype that combines wearable sensing and camera-derived vitals
Achieved live telemetry over both ESP-NOW and UDP
Implemented adaptive BPM detection logic with timeout and refresh behavior
Added practical on-device feedback (OLED) for field usability
Structured the project into modular sketches and communication components

What we learned

Embedded health sensing requires careful filtering, thresholds, and calibration
Motion strongly affects optical pulse quality, so cross-checking with activity data is important
Robust rescue-oriented systems need graceful failure handling (timeouts, reconnects, fallback states)
Communication design (packet format, retry behavior, encryption choices) is as important as sensing logic
Iterative testing with real hardware is essential for stable performance