Bastille, a San Francisco-based startup, says its software-defined radio sensors will be able to protect enterprises from security threats to their networks.

The company’s patented technology pairs software-defined radio sensors with machine learning to identify abnormal events on the wireless network regardless of whether it is WiFi, Bluetooth, cellular, or another Internet of Things (IoT) protocol.

Bastille says it uses three different types of technologies to scan the spectrum from 60 MHz to 6 GHz. First, it uses collaborative bandit sensing, which scans the spectrum for emitters and threats using a Bandit prediction algorithm and machine learning. Second, it uses Bayesian device fingerprinting, which detects and identifies devices in the enterprise’s airspace. And third, it uses Distributed Tomographic Localization, which provides position information of all emitters in a corporate airspace.

According to Bastille CEO Chris Risley, the company can determine if a device on the network is acting abnormally and identify threats or attacks that might be occurring. For example, it can tell if there’s a rogue WiFi hotspot, a surveillance device, or even an unapproved wireless peripheral like a keyboard or a camera on the network.

Because Bastille’s sensors are based upon software-defined radios, it can adjust or upgrade protocols without having to replace radios or chip sets. “We can deploy these sensors, and if a new protocol comes out, we just push a new software update,” said Bob Baxley, chief engineer at Bastille.

The company’s security-as-a-service business model provides companies with the software-defined radio sensors for free but charges for security updates and alerts that are incorporated into the company’s security information management (SIM) system.