Active & Reactive Threat Protection For Our Warfighters: Electronic Warfare & Edge AI

Case Study: Machine Learning in safety critical environments on NATO edge hardware

improvised explosive devices (IEDs) are homemade bombs frequently employed by guerrillas, insurgents, and other non-state actors as a crude yet effective weapon against conventional military forces. These devices are designed to destroy, incapacitate, harass, or distract, and can range from small pipe bombs to sophisticated devices capable of causing significant damage and loss of life.

Roadside IEDs can be initiated through various methods, falling into two main categories: command-initiated and autonomously initiated. Command-initiated IEDs are most common and are detonated through human interaction with a triggering mechanism, often a receiver that triggers detonation via a wireless signal. These signals can be jammed, protecting those in the area and preventing detonation.

Designing effective radio frequency (RF) waveforms to inhibit a range of threat devices, such as those used in IEDs and Unmanned Aerial Systems (UAS), is a complex and labor-intensive process. This typically involves RF surveys of the target environment, followed by the collection of requirements for threat profiles, including devices, transmitter ranges, network technologies, and protection area size. This has traditionally been the domain of manual profiling, using highly skilled manual waveform design by electronic countermeasures (ECM) units. This entire process is time-consuming, relies heavily on expert human operators, and can significantly delay the deployment of critical force protection measures.

Polygeist has pioneered the application of machine learning (ML) to revolutionize active waveform generation. Our innovative approach, powered by sophisticated models has been able to deliver optimized jamming at unprecedented speeds, far surpassing traditional human-led methods.

We’ve also innovated in deployment context, starting on large laboratory workstation devices, we’ve been able to shrink this capability to dedicated hardware, allowing low-powered devices to be deployed on the battlefield.

The details of this technology, and our partners are currently under embargo, but we are proud to report that Polygeist technology is being deployed into the NATO supply chain. We’re demonstrating that cutting edge ML can be developed and deployed in low-compute, safety critical arenas.