Transformer-based radio signal classifier for noisy RF
This technology uses a two-stage artificial intelligence system to analyze and interpret radio signals. A first neural network processes an incoming radio signal and produces an initial analysis, which is then fed—along with the original signal—into a 'signal transformer' that reshapes the signal in ways informed by that first analysis. A second neural network then processes this transformed version to extract detailed information about the signal. The result is a smarter, more adaptive approach to understanding what a radio signal contains or represents compared to traditional single-pass processing.
What you could build
A software or firmware module for wireless receivers—such as those used in spectrum monitoring, cognitive radio, or electronic warfare systems—that improves signal identification and decoding accuracy. Likely buyers include defense contractors building signals intelligence (SIGINT) platforms, telecom equipment vendors, and spectrum management companies.
Who in Virginia should care
Defense primes and SIGINT contractors concentrated in Northern Virginia (Booz Allen, Leidos, SAIC, CACI) would have direct operational interest in improved radio signal classification for electronic warfare and intelligence applications.
Readiness: Lab validated
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
- —
- Granted
- May 21, 2019
- Status
- Granted patent
- Patent number
- 102831
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.