How can wireless operators predict interference before deploying equipment? Federated Wireless says its latest Spectrum AI platform can emulate radio propagation, coverage, and network capacity earlier than antennas release. The company designed the platform to support mobile operators, cable providers, and wireless internet service providers make better use of shared spectrum.
Federated Wireless introduced the general availability of Spectrum AI following the release of its Adaptive Network Planner in advance this year. Together, the tools give network engineers in detailed view of potential deployment sites and anticipated network performance.
Using Physical AI to Model Wireless Networks
Federated Wireless explains Spectrum AI as a physical AI platform formed particularly for shared spectrum planning and coordination. It helps Citizens Broadband Radio Service Priority Access License and General Authorized Access customers, along side 6 GHz networks subject to incumbent-protection rules.
Unlike AI system that automate network workflows above the physical layer, Spectrum AI functions on the radio propagation level. It models how electromagnetic signals move through buildings, trees, terrain, and other physical structures.
Engineers can use its 3D modeling capabilities to analyze rooftop antenna locations, adjust antenna height, and evaluate how small placement change could impact on signal quality.
The platform uses of NVIDIA’s open-source Sionna RT ray-tracing library within an Amazon Web Services environment that integrates Amazon Bedrock. As per Federated Wireless CEO Iyad Tarazi, the system can carry out billions of simulations for individual cell sites.
Anticipating Interference Before Deployment
Federated Wireless designed Spectrum AI partly in response to interference and spectrum restrictions impacting operators near protected or congested environments.
Tarazi pointed to Amplex Internet, an Ohio-based wireless internet service supplier going through interference from higher-powered Canadian wireless networks. He stated Spectrum AI could support operators model those conditions earlier than deployment instead of assigning a channel and waiting to see whether or not interference happens.
The system can evaluate client speeds, coverage, capacity, and service economics before equipment is installed. Federated Wireless also claims its simulations can produce predictions comparable to physical drive testing.
One operator reportedly informed the company that it planned to replace drive testing in one state with Spectrum AI simulations.
Increasing Usable Spectrum Capacity
Federated Wireless says about 25% of sites functioned by using massive providers face limited intended to prevent interference. These limits can lessen the speed and capacity operators can offer.
By simulating network behavior and monitoring interference, the company claims it could remove up to 99% of certain limitations. Field deployments have also shown up to 5 times more greater network capacity and as much as 50% more usable spectrum capability, as per Federated Wireless.
The company also mentioned an improvement of about 20 decibels in interference-coordination accuracy across live shared-spectrum environments.
Tarazi said network simulations that previously needed two weeks and a server farm can now run in seconds. Moreover, the system relies upon on persisted updated environmental and field data. Federated Wireless spent two years assembling information about buildings, structures, vegetation, and other factors that influence radio propagation.
