Back to the portfolio

Adaptive learned MIMO encoder for 5G and satellite radios

Communications & RFComputing, Software & AIElectronics & SemiconductorsLab validated

This technology uses machine learning to dramatically improve how wireless signals are sent and received across multiple antennas simultaneously — a setup called MIMO, which is central to modern 5G and Wi-Fi systems. The system trains a neural network to understand the specific characteristics of a wireless channel in real time, using feedback from the receiver to continuously refine how data is encoded and transmitted. Instead of relying on fixed, rule-based signal processing, the AI adapts the transmission strategy on the fly, choosing optimal signal patterns (constellations) for each antenna based on actual channel conditions. The result is a smarter, more efficient wireless link that can recover more of the original information even in noisy or complex environments.

What you could build

A software-defined radio module or chipset firmware layer that embeds learned MIMO encoding/decoding, targeting wireless infrastructure OEMs, 5G base station vendors, and satellite communication equipment makers who need higher spectral efficiency without adding antenna hardware.

Who in Virginia should care

Northern Virginia's concentration of telecom infrastructure operators, federal contractors running secure communications networks, and data center operators deploying private 5G would be natural early evaluators or licensees.

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, Tugba Erpek
Granted
January 2, 2024
Status
Granted patent
Patent number
11863258

Ready to talk?

Virginia Tech Intellectual Properties handles licensing for this technology.

VTIP contact coming shortly

Prosim summaries are generated from public patent text and are not legal advice.