AI-adaptive MIMO encoding IP for real-world wireless channels
This technology uses machine learning to train wireless transmitters and receivers to work together more effectively across MIMO channels — the multi-antenna systems that underpin modern 5G and Wi-Fi. Instead of using fixed, hand-engineered signal encoding and decoding rules, both ends of the communication link are represented as neural networks that learn jointly how to send and receive information. The system feeds data through a simulated channel model, measures how much information is lost or distorted, and updates the neural networks to minimize that error over many training cycles. The result is a communication system that can adapt to complex or unusual channel conditions that traditional signal processing methods handle poorly.
What you could build
An AI-driven physical-layer design toolkit or chip IP block for next-generation wireless equipment — sold or licensed to semiconductor companies, base station vendors, or satellite communication hardware makers who need adaptive, high-throughput MIMO performance beyond what classical signal processing delivers.
Who in Virginia should care
Northern Virginia's dense cluster of defense contractors, satellite communications firms (e.g., those supporting LEO constellations), and federal telecom R&D agencies (DARPA, NSA, NRL) would have direct interest in adaptive MIMO physical-layer research.
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, Tugba Erpek
- Filed
- Patent pending — filed June 13, 2025
- Status
- Application
- Publication number
- US20250373351A1
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.