KofKaN AI Health & Environmental Monitoring System
Advanced Edge-AI & Multi-Sensor IoT Healthcare Hardware
Overview
Designed from circuit schematics to a ruggedized physical enclosure, this project integrates high-performance microcontrollers with multi-sensor arrays to safeguard indoor environments. The architecture pairs an ESP32-P4 main processing engine—handling local neural network inference and high-speed image processing via a Raspberry Pi v2.1 (IMX219) camera—with an auxiliary ESP32-S3 sentry node for pre-event video buffering. A custom-built mobile application dashboard provides real-time telemetry updates for carbon monoxide, oxygen, humidity, dust density, and electromagnetic fields.
System capabilities
Edge-AI Vision Processing
Runs local neural network inferencing on high-performance RISC-V cores using MIPI-CSI camera feeds.
Multi-Sensor Telemetry
Real-time tracking of CO, Oxygen, Humidity, Dust Density, and EMF metrics.
Pre-Event Buffering
Maintains a rolling 3-second video buffer via auxiliary node to capture context leading up to incidents.
How it works
- 01Auxiliary ESP32-S3 node continuously captures 3-second rolling video buffers in PSRAM.
- 02Environmental and gas sensors transmit telemetry data over local IoT protocols.
- 03A hardwired GPIO interrupt triggers the main ESP32-P4 processor instantly upon a fall event.
- 04ESP32-P4 executes heavy edge-AI pose inference via ESP-DL and local vision models.
- 05Custom Android dashboard displays live sensor metrics and room health status.
Applications
- Elderly fall detection and emergency response
- Continuous indoor air quality and gas monitoring
- Smart home and assisted living telemetry
- Remote environmental safety logging