edgeGuard is a compact, helmet-mountable edge-AI platform designed for first responders operating in hazardous, low-visibility, and connectivity-constrained environments. Based on the accompanying paper, the platform combines onboard vision, environmental sensing, motion sensing, audio capture, wireless communication, and local AI inference in a board footprint of approximately 4 cm x 2.5 cm.
This open-source package currently focuses on the hardware design files and supporting documentation for the EdgeGuard platform.
- Ultra-compact edge-AI hardware built around the GAP9 AI MCU for near-sensor processing.
- Multimodal sensing with RGB vision, gas sensing, pressure, humidity, temperature, IMU, and microphone inputs.
- Wireless-ready architecture using the nRF5340 for low-power coordination and BLE 5.4 connectivity.
- Emergency-response focus with onboard pipelines for fire, smoke, person, and hazardous-gas perception.
- Long battery-backed operation reported in the paper at more than 10 hours from a 600 mAh LiPo battery.
- GAP9 AI MCU with a 9-core RISC-V compute cluster and NE16 accelerator.
- External memory including 512 Mbit Octal-SPI Flash and 512 Mbit PSRAM.
- Nordic nRF5340 dual-core wireless SoC for BLE 5.4 connectivity and low-power coordination.
- Nordic nPM1300 PMIC for battery charging and system power management.
- FT4232HL FTDI interface for programming, debugging, and USB data forwarding.
- Himax HM0360 VGA image sensor with MIPI CSI-2 interface.
- Bosch BME688 for temperature, humidity, pressure, and gas sensing.
- Dedicated CO and CH4 sensing channels for hazardous-environment monitoring.
- ST LSM6DSO 6-DoF IMU for motion context.
- Digital PDM microphone for ambient audio capture.
The system was tested in two representative onboard inference pipelines:
- A lightweight YOLO-based detector for fire, smoke, and person detection reaching 0.80 mAP@50.
- An SVM-based environmental-sensing pipeline for hazardous gases and flammables reaching nearly 92% accuracy.
References to the publication will be shortly added.
-
Hardware/Contains the Altium design sources for the EdgeGuard PCB. -
Hardware/edgeguard_mainboard/Contains the main board project files, libraries, schematic sheets, harness files, and PCB layout. -
Firmware/Reserved for future open-source release of firmware, embedded software, and deployment instructions. -
Documents/Contains supporting repository documentation. -
Documents/Schematics/Contains the exported schematic PDF for quick review without Altium. -
Documents/Pictures/Images/Contains images used in the documentation, including the PCB figure shown above.
This project is open source.
The files within the following folder (and subfolders) are released under Solderpad v0.51 (SHL-0.51) (see LICENSE.txt):
Hardware/
The files within the following folder (and subfolders) are released under Creative Commons Attribution 4.0 International (CC-BY-4.0) (see Documents/LICENSE):
Documents/
At this stage, Firmware/ is documentation-only and should be treated as part of the repository metadata until source files are added.