On July 13, 2023, the Federal Communications Commission (FCC or Commission) released a Draft Notice of Inquiry (Draft NOI or Draft) that would seek input on leveraging new technologies to collect and analyze data on non-Federal spectrum usage, including through machine learning (ML) and artificial intelligence (AI). As the radiofrequency environment becomes more congested, the FCC seeks new tools to better identify opportunities to facilitate more intensive spectrum use on a “cost-effective, accurate, scalable, and actionable manner.”
The Draft NOI arises on the heels of Chairwoman Rosenworcel’s recent AI workshop, signaling the Commission’s increasing focus on AI’s capabilities and potential. In this case, the Commission hopes to build on its recent efforts to incorporate next-generation techniques and data-driven analytics into its spectrum management toolkit. Citing its recent policy statement regarding the expectation that receivers will operate resiliently in congested spectrum bands, the Commission suggests that its findings in this new proceeding may impact how the agency approaches spectrum use assessments in future proceedings. The Draft NOI will be discussed and voted upon at the August 3, 2023 FCC Open Meeting. Below we provide a high-level summary of the Draft NOI’s key points.
* * *
The Draft NOI would seek information on three major factors that impact the collection and analysis of spectrum usage data: (1) the appropriate definition of “spectrum usage”; (2) uniformity of band requirements and measurements; and (3) data collection techniques. It also would seek comment on associated challenges—including cost, band-specific-considerations, and the need for uniform metrics and data collection requirements—as well as next steps the Commission might take to further its inquiry.
Definition of Spectrum Usage
The Draft NOI would seek comment on how to define spectrum usage, including appropriate terms and definitions. The Draft considers, for example, the definitions of “band occupancy” used in technical studies by the National Telecommunications and Information Administration (NTIA) and National Institute of Standards and Technology (NIST), as well as “spectrum occupancy” cited in reports by the International Telecommunication Union (ITU). The FCC would further seek comment on whether to examine spectrum usage through discrete components, such as geographic usage, frequency usage, and time usage, or by other categories such as design techniques or engineering metrics (e.g., throughout, population served, or the number or density of end-user devices).
Uniformity of Band Requirements and Measurements
The Draft NOI acknowledges that spectrum bands do not have uniform service requirements, operational systems, or technical characteristics. As a result, past spectrum studies have concluded that there is no “one-size-fits-all” approach for measuring spectrum usage. The Draft NOI would ask whether commenters agree with this assessment, and if so, how the agency should account for spectrum usage variances. For example, should subscriber-based services be evaluated differently from privately controlled systems? Should underlying reliability or service requirements be taken into account? Should unlicensed use be factored in, if at all? Should usage metrics be adjusted based on the geographic area over which a license is used?
Additionally, the Draft NOI would ask for comment on the accuracy of representative sampling for measuring usage, and if so, what geographical areas can “serve as a reliable proxy for nationwide conclusions?” More broadly, the FCC would seek comment on whether there are other appropriate metrics that the agency should prioritize in studying spectrum usage.
Data Collection
Finally, the Draft NOI focuses on the actual process of collecting spectrum usage data and would seek comment on its existing spectrum usage data collection tools, including its licensing databases: the Universal Licensing System (ULS) and International Communications Filing System (ICFS). The Draft NOI would invite feedback on helpful additional data to collect and would seek guidance on where that data could be obtained. Specifically, the Commission notes that past commenters have suggested utilizing crowdsourcing, external data sources, modeling, and direct observation as methods for accurate and robust data collection. The Draft NOI is interested in comments about how existing spectrum measurement tools can be best utilized and what emerging tools might be useful to collect data accurately and efficiently.
The FCC identifies cost, standardization, and technical accuracy as potential obstacles to accurate data collection and would seek comment on how to mitigate these challenges. The Draft NOI proposes several potential solutions such as cabining the scope of data collected to certain areas that serve as proxies for the nation, requiring standardized data formats and methodologies, and utilizing AI and ML to extrapolate broad conclusions on spectrum usage from smaller data sets. The FCC would also ask for comment on additional obstacles to data collection.
The Draft NOI includes an acknowledgement of the privacy and cybersecurity concerns with data collection generally and would request input on how existing privacy laws might apply if the Commission conducts this type of data collection.
Conclusion
The Draft NOI broadly outlines numerous questions regarding existing data collection and analytics practices for studying non-Federal spectrum usage and highlights some of the Commission’s goals for improvement. While the Draft NOI only briefly discusses emerging technologies, the item no doubt marks the beginning of the FCC’s interest in leveraging AI and ML analytics to improve its spectrum management practices.
This item has not yet been approved by the Commission. If it receives a passing vote, it will be published in the Federal Register and comments are scheduled to be due 60 days later.
[View source.]
