Edge AI compression engine for RF and wireless signals
This technology uses a pair of neural networks — one to compress a radio frequency signal and one to reconstruct it — trained together to find the smallest possible representation of an RF signal that still allows accurate recovery. The training process automatically balances two competing goals: keeping the compressed signal as small as possible while minimizing the error when the signal is rebuilt. Once trained, the encoder can be deployed on a transmitter or sensor and the decoder on a receiver, enabling efficient RF signal transmission or storage. This is essentially learned compression, applied to wireless signals the same way modern image codecs apply learned compression to photos.
What you could build
A software module or chip-level IP block that compresses RF signals at the edge — on base stations, IoT sensors, or spectrum monitoring hardware — before transmission or logging, reducing bandwidth and storage costs. Telecom equipment vendors, defense electronics integrators, and spectrum analytics companies would be the buyers.
Who in Virginia should care
Northern Virginia defense and intelligence contractors (Leidos, Booz Allen, SAIC, Noblis) working on spectrum management, SIGINT, and 5G infrastructure modernization would have direct use cases.
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
- Timothy James O`Shea
- Granted
- February 25, 2020
- Status
- Granted patent
- Patent number
- 10572830
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.