Threat Detection

A Single-Node Detection

Soundryx sensors identify threats by their sound waves.

Each sensor has an array of microphones that listens to ambient sounds and also determines the direction that different sounds arrive from.

An on-sensor machine learning (ML) model analyzes the data from the microphone array for evidence of drone aircraft, gunshots, and other potential threats.

The actual audio recorded by the microphone array is discarded. The ML model's threat analysis is what gets sent to the backend.

Network Detection

When a sensor detects a potential threat using its onboard machine learning model, it sends a short message via LTE, Wi-Fi, or wired Ethernet to a Soundryx backend server.

The backend receives alerts from every sensor in the network, and can combine alerts from multiple sensors to calculate the speed and direction of detected threats.