Independent Calculation of Move Lists for Incumbent Protection in a Multi-SAS Shared Spectrum Environment.

In the 3.5 GHz Citizens Broadband Radio Service (CBRS), secondary users are managed by spectrum access systems (SASs) to protect incumbents from interference. Current practice requires each SAS to exchange detailed user information with other SASs, and to use a common algorithm to suspend transmissions so that an aggregate interference percentile is below a predefined threshold. We propose a simplified method that utilizes a tight bound on the aggregate interference distribution. Simulation results show that the proposed approach trades off a marginal reduction in spectral efficiency to greatly simplify incumbent protection procedure, allowing each SAS to independently manage its users.

Deterministic Move Lists for Federal Incumbent Protection in the CBRS Band.

The 3.5 GHz citizens broadband radio service (CBRS) band in the U.S. is a key portion of mid-band spectrum shared between commercial operators and existing federal and non-federal incumbents. To protect the federal incumbents from harmful interference, a spectrum access system (SAS) is required to use a common, standardized algorithm, called the move list algorithm, to suspend transmissions of some CBRS devices (CBSDs) on channels in which the incumbent becomes active. However, the current reference move list implementation used for SAS testing is non-deterministic in that it uses a Monte Carlo estimate of the 95th percentile of the aggregate interference from CBSDs to the incumbent. This leads to uncertainty in move list results and in the aggregate interference check of the test. This paper uses upper and lower bounds on the aggregate interference distribution to compute deterministic move lists. These include the reference move list used by the testing system and an operational move list used by the SAS itself. We evaluate the performance of the proposed deterministic move lists using reference implementations of the standards and simulated CBSD deployments in the vicinity of federal incumbent dynamic protection areas.

Deep Learning Classification of 3.5 GHz Band Spectrograms with Applications to Spectrum Sensing.

In the United States, the Federal Communications Commission has adopted rules permitting commercial wireless networks to share spectrum with federal incumbents in the 3.5 GHz Citizens Broadband Radio Service band. These rules require commercial systems to vacate the band when sensors detect radars operated by the U.S. military; a key example being the SPN-43 air traffic control radar. Such sensors require highly-accurate detection algorithms to meet their operating requirements. In this paper, using a library of over 14,000 3.5 GHz band spectrograms collected by a recent measurement campaign, we investigate the performance of thirteen methods for SPN-43 radar detection. Namely, we compare classical methods from signal detection theory and machine learning to several deep learning architectures. We demonstrate that machine learning algorithms appreciably outperform classical signal detection methods. Specifically, we find that a three-layer convolutional neural network offers a superior tradeoff between accuracy and computational complexity. Last, we apply this three-layer network to generate descriptive statistics for the full 3.5 GHz spectrogram library. Our findings highlight potential weaknesses of classical methods and strengths of modern machine learning algorithms for radar detection in the 3.5 GHz band.

3.5 GHz Environmental Sensing Capability Detection Thresholds and Deployment.

Spectrum sharing in the 3.5 GHz band between commercial and government users along U.S. coastal areas depends on an environmental sensing capability (ESC)-that is, a network of radio frequency sensors and a decision system-to detect the presence of incumbent shipborne radar systems and trigger protective measures, as needed. It is well known that the sensitivity of these sensors depends on the aggregate interference generated by commercial systems to the incumbent radar receivers, but to date no comprehensive study has been made of the aggregate interference in realistic scenarios and its impact on the requirement for detection of the radar signal. This paper presents systematic methods for determining the placement of ESC sensors and their detection thresholds to adequately protect incumbent shipborne radar systems from harmful interference. Using terrain-based propagation models and a population-based deployment model, the analysis finds the offshore distances at which protection must be triggered and relates these to the detection levels of coastline sensors. We further show that sensor placement is a form of the well-known set cover problem, which has been shown to be NP-complete, and demonstrate practical solutions achieved with a greedy algorithm. Results show detection thresholds to be as much as 22 dB lower than required by current industry standards. The methodology and results presented in this paper can be used by ESC operators for planning and deployment of sensors and by regulators for testing sensor performance.