Artificial intelligence, machine learning and similar technologies (collectively AI) are poised to transform industries, institutions and our daily lives. The past year alone has shown how diversely these technologies will change the way we communicate, function and receive and understand information. As government actors grapple with whether and how to regulate AI, they are also considering how to use these technologies to strengthen their licensing and oversight obligations. That is exactly what the FCC is doing. Last week, the FCC released a draft Investigation Notice (NOI) on how AI can be used to increase its knowledge and analytical capabilities related to commercial uses of the radio frequency spectrum. Why? Publicly available commercial spectrum usage data remains scarce; Therefore, the FCC relies on third-party studies to review new spectrum usage.
This proceeding is not about the underlying spectrum policies or service rules of the Commission. However, it will advance the FCC’s 2023 Spectrum Policy Statement, in which it advocates modern technologies to increase the FCC’s knowledge of commercial spectrum use in a “cost-effective, accurate, scalable, and actionable manner.”
Current data collection methods for spectrum monitoring
It is a challenge to get spectrum usage data in real-time. For example, the Universal Licensing System, International Communications Filing System, and other FCC databases lack real-time information on licensed spectrum usage. White space databases and automated frequency coordination systems track available spectrum for secondary use, but do not assign users to specific channels. Recognizing these shortcomings, the FCC authorizes third-party administrators to develop and maintain spectrum access systems to monitor and coordinate in-band sharing. The FCC also conducts speed and ride tests by cell phone operators to collect data on network coverage and broadband speeds. Despite these efforts, real-time data on spectrum usage is not available for almost all spectrum bands.
Other U.S. agencies and international bodies have considered or are considering similar options. For example, the National Telecommunications and Information Administration examines federal frequency data as part of its Spectrum Analysis Program and the Institute for Telecommunication Sciences recently examined the Citizens Broadband Radio Service. The National Science Foundation (NSF) and National Institute of Standards and Technology (NIST) have addressed the benefits and challenges of spectrum data with recent initiatives, including the NIST assessment of spectrum usage during the COVID-19 pandemic.
Several international bodies also monitor frequency usage. The International Telecommunication Union has highlighted the role of spectrum monitoring in effective spectrum management and interference remediation. Member State administrations operate the International Monitoring System, which has some 400 stations in 81 countries, collecting data, sending reports and publishing summaries of spectrum usage. The UK deploys spectrum detectors across the country to measure and understand spectrum usage in specific areas. The Canadian Communications Research Center has further developed a prototype system for visualizing spectrum data.
Taken together, these efforts underscore the importance of data-driven spectrum management at home and abroad.
In the draft NOI, the FCC is asking for input in four areas.
Define Spectrum Usage. How should spectrum usage be defined? What are the pros and cons of previous initiatives to define, understand and measure spectrum usage? Should the FCC separate spectrum usage into components such as geographic usage, spectrum usage, and time usage? Besides the mere presence of a signal of a certain strength, what other RF technical metrics could assess spectrum usage? Can spectrum usage metrics be combined to generate a holistic understanding of the radio frequency landscape?
Band specific issues. What are best practices, operational considerations and technical parameters that might correspond to different aspects of spectrum usage in different radio services? How should the FCC prioritize data collection when each issue or volume has its own unique set of challenges? Any comments on the 2016 NSF workshop conclusions that support measurement of traditional fixed and mobile terrestrial transmitters and sub-6GHz bands? How should the Commission tailor its research techniques based on licensing and usage characteristics for specific bands?
data meetings. What data sources could help understanding spectrum usage? What are existing data sources? What are the data-related challenges, such as cost and effort, standardization, and technical accuracy? What are the pros and cons of different data collection methods, including crowdsourcing, external data sources, modeling, and direct observation?
Other matters. What are other practical, technical, and legal concerns about spectrum use, including privacy and security, FCC legal powers, and digital equity and inclusion considerations? How should the FCC facilitate and incentivize data sharing? Should it launch a field monitoring pilot program in the near future to investigate non-federal spectrum use? According to which criteria should it be checked how other interest groups (authorities, universities, private institutions) can support their research and reporting? In the longer term, should it issue non-binding guidelines that might set out best practices for assessing spectrum use?
The FCC will consider the draft NOI at its public meeting on August 3, 2023.
* * *
Many thanks to Summer Associates Jordyn Johnson, Isabelle Dean and Ryan Campbell for their valuable contributions to this publication.