Transformer-based AI engine for radio signal classification
This technology uses a layered AI architecture to analyze and interpret radio signals more accurately than conventional methods. A first neural network processes a raw radio signal and produces preliminary data, which is then fed alongside the original signal into a 'signal transformer' — an AI model adapted from the same architectural breakthroughs behind large language models. The transformer applies learned mathematical operations to produce a refined version of the signal, which a second neural network then interprets to extract meaningful information such as signal type, source, or modulation scheme. The result is a more capable, data-driven radio signal recognition system that can learn complex patterns without relying solely on hand-engineered signal processing rules.
What you could build
A software-defined radio intelligence engine, sold as an embedded module or cloud API, that automatically classifies, characterizes, and decodes unknown radio signals — bought by spectrum monitoring firms, defense electronics integrators, and wireless network operators managing complex RF environments.
Who in Virginia should care
Northern Virginia and the Hampton Roads corridor host major defense contractors (Leidos, SAIC, Raytheon), NSA-adjacent signals intelligence programs, and DoD spectrum management offices that are direct natural buyers for RF signal intelligence technology.
Readiness: Concept
Concept — described but not yet demonstrated. Lab validated — supported by experimental results in the patent. Prototype likely — the text describes a built, working embodiment.
Readiness is inferred from the patent text, not from a lab visit.
The record
- Inventors
- Timothy James O`Shea
- Filed
- Patent pending — filed August 9, 2024
- Status
- Application
- Publication number
- US20250045581A1
Ready to talk?
Virginia Tech Intellectual Properties handles licensing for this technology.
Prosim summaries are generated from public patent text and are not legal advice.