ĢƵ Allen’s prototype system first digitizes the signals on the shared spectrum then uses artificial intelligence (AI) to identify potential conflicts. The system then determines what measures are needed to deconflict the signals and sends those to the commercial 5G station to implement.
At the heart of the solution is ĢƵ Allen’s patented R.AI.DIO® capability that brings AI to radio signal processing. R.AI.DIO® signal processing capability—which ĢƵ Allen developed as an internal technology investment and has leveraged for other government clients—has a wide range of applications. These include making communications more secure and resilient, detecting rogue communication signals, and providing early warning of adversary jamming efforts. R.AI.DIO® processing can also be used on large unprocessed radio frequency (RF) datasets to quickly triage regions of interest and enable faster signal discovery.
ĢƵ Allen’s approach substantially accelerates the decision analytics involved in spectrum sharing and coexistence solutions.
Here are the four basic steps in ĢƵ Allen’s AI-enabled spectrum sharing and coexistence solution:
- A high-dynamic-range software-defined radio continuously digitizes all of the signals on the shared channel and removes any noise.
- ĢƵ Allen’s R.AI.DIO® neural networks use AI to rapidly identify the commercial 5G signals and military airborne radar signals that might conflict.
- Potentially conflicting 5G and military signals are passed to a “command word generator,” which decides which, if any, interference-mitigation measures should be taken. These measures can include shifting the 5G signals to another frequency, lowering their power level, or applying other mitigation measures in the 5G system.
- The system interfaces with the relevant commercial 5G base station, which can then automatically execute any mitigation measures needed to protect the radar.
ĢƵ Allen began the project in February 2021 and is implementing a 15-month proof-of-concept phase. This includes developing, testing, and optimizing the software and hardware components and then testing the system as a whole in ĢƵ Allen’s RF laboratories.
The testing takes place in a mixed-signal environment that simulates the effects of terrain, land clutter, buildings, atmospheric obstructions, and other real-world conditions that might affect RF performance.
Following the proof of concept, ĢƵ Allen will conduct incremental testing and improvements to the system for 1 year at its own labs and at an Air Force electromagnetic compatibility facility. This lab testing will be followed by at least 1 year of field testing at Hill Air Force Base in Utah, where the airborne radars involved in the project are based.