Neural network MIMO encoder/decoder for 5G and Wi-Fi 7
This technology uses neural networks to automatically learn the best way to encode and decode wireless signals sent between devices with multiple antennas — a configuration called MIMO, which is the backbone of modern Wi-Fi and cellular networks. Instead of engineers hand-designing the encoding rules, a machine learning model trains end-to-end by simulating the wireless channel, learning simultaneously how to pack data in at the transmitter and unpack it accurately at the receiver. The system adapts to different channel conditions by modeling their effects during training, producing encoding strategies that can outperform conventional designs in noisy or complex radio environments. The result is a jointly optimized transmitter-receiver pair that can generalize across varying real-world wireless conditions.
What you could build
A software-defined radio (SDR) or chipset firmware layer that replaces conventional MIMO signal processing with a trained neural network encoder/decoder, sold or licensed to wireless chipset makers and basestation OEMs seeking performance gains in 5G/6G or Wi-Fi 7 deployments.
Who in Virginia should care
Northern Virginia's dense concentration of wireless infrastructure operators, defense communications contractors (e.g., SAIC, Leidos, Noblis), and federal spectrum users would be natural early customers or partners for field-trial evaluation.
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 12, 2021
- Status
- Granted patent
- Patent number
- 10892806
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.