ruvnet

RuView

AIESP32WiFi SensingEdge AINeural NetworksCSI

123

// summary

RuView is an edge-based sensing platform that utilizes WiFi Channel State Information (CSI) to detect human presence, vital signs, and activities without cameras or wearables. The system leverages ESP32 hardware and spiking neural networks to perform real-time spatial intelligence, including through-wall monitoring and pose estimation. It operates entirely locally without cloud dependencies, offering privacy-focused tracking and environment mapping through radio signal analysis.

// technical analysis

RuView is a specialized WiFi sensing platform that leverages Channel State Information (CSI) from low-cost ESP32 hardware to perform spatial intelligence tasks without cameras or wearables. By analyzing radio wave disturbances, the system enables contactless monitoring of human presence, vital signs, and activity recognition through walls. The architecture prioritizes edge-native processing, utilizing spiking neural networks and multi-frequency mesh scanning to adapt to environments in under 30 seconds while maintaining privacy through local, cloud-free execution.

// key highlights

Enables contactless vital sign monitoring, including breathing and heart rate, by analyzing sub-millimeter radio wave disturbances.

Provides high-accuracy presence detection and occupancy counting that functions through walls and in total darkness.

Supports 17-keypoint pose estimation using the WiFlow architecture, trained via camera-supervised or camera-free methods.

Utilizes a multi-frequency mesh approach that turns neighboring WiFi routers into passive radar illuminators to increase sensing bandwidth.

Features a persistent edge-AI pipeline that stores sensing events as searchable vectors, enabling anomaly detection and environment fingerprinting.

Implements cryptographic attestation for all measurements, ensuring data integrity and security without requiring cloud connectivity.

// use cases

Contactless vital sign monitoring including breathing and heart rate detection

Through-wall human presence detection, occupancy counting, and activity recognition

Real-time 3D point cloud generation and pose estimation using WiFi CSI and sensor fusion

// getting started

To begin, you can either run the Docker image for simulated data evaluation or flash the provided firmware onto an ESP32-S3 board for live sensing. Once the hardware is provisioned with your WiFi credentials, you can use the provided Node.js scripts to perform tasks like RF room scanning, person counting, or real-time pose estimation. For advanced persistent storage and AI integration, the system supports optional Cognitum Seed hardware.