Spectrum Access System (SAS)

The SAS maintains a secure connection to the Environmental Sensing Capability (ESC) network to detect federal radar operations. Upon detection, the SAS calculates a Dynamic Protection Area (DPA) and immediately instructs all Citizens Broadband Radio Service Devices (CBSDs) within that zone to cease transmission on the affected channels. This process must complete within 60 seconds to meet regulatory requirements.

• ESC Sensors: Detect specific federal radar signatures near coastlines.
• DPA Activation: Moves dynamically with maritime radar systems.
• Grandfathered Wireless Protection: Protects fixed satellite service earth stations.

Every CBSD must register with the SAS, providing precise geolocation, antenna height, and requested power levels. The SAS uses a propagation model (e.g., Irregular Terrain Model) to calculate aggregate interference. It ensures that the cumulative transmission from all secondary users does not exceed the -80 dBm/m² interference threshold at the boundary of any protected contour.

• Professional Installation: Required for Category B CBSDs to ensure location accuracy.
• Propagation Modeling: Accounts for terrain clutter and morphology.
• Aggregate Interference Control: Manages the sum of all emissions, not just individual links.

This is the core authorization logic. A CBSD sends a Spectrum Inquiry message, and the SAS responds with a Spectrum Grant or denial. The engine applies the two-tier commercial authorization logic: Priority Access Licenses (PALs) receive interference protection from General Authorized Access (GAA) users, but GAA users must not interfere with each other. The engine uses a contiguous channel allocation algorithm to maximize bandwidth availability.

• Heartbeat Protocol: CBSDs must send periodic heartbeats to maintain the grant.
• Grant Suspension: The SAS can suspend a grant instantly if an incumbent appears.
• Channel Assignment: Assigns specific frequency ranges (e.g., 3550-3560 MHz).

Multiple SAS administrators (e.g., Google, Federated Wireless) can operate simultaneously in the same geographic area. The SAS-to-SAS Coordination Protocol ensures they share information about registered CBSDs and active grants. This prevents a 'hidden node' problem where a CBSD registered with one SAS causes interference to a protected entity managed by another SAS. The protocol synchronizes the Coordination Area records daily.

• Full Activity Dump: Daily synchronization of all grant records.
• Escalation Mechanism: Resolves conflicts between competing SAS administrators.
• Interoperability Testing: Mandated by the WinnForum standards body.

The SAS maintains a comprehensive database that is the authoritative record of spectrum rights. It stores the geographic boundaries of PAL licenses, the exclusion zones for federal incumbents, and the historical logs of all spectrum grants. This repository is used for post-hoc interference analysis and regulatory compliance auditing. It ensures that a CBSD cannot claim ignorance of a nearby protected entity.

• PAL Database: Defines the census tracts where licenses are valid.
• Exclusion Zone Maps: Hard-coded boundaries where no CBSD operation is allowed.
• Audit Trail: Immutable log of all authorization decisions.

The machine learning paradigm increasingly used to optimize SAS decision logic. In an RL framework, the SAS acts as an agent that learns an optimal channel allocation policy by interacting with the RF environment. The agent receives rewards for maximizing spectrum utilization and penalties for causing interference. Algorithms like Deep Q-Networks (DQN) and Proximal Policy Optimization (PPO) enable the SAS to handle the high-dimensional state space of real-world spectrum occupancy, moving beyond rigid rule-based heuristics.

An integrated spatial-spectral database that serves as the SAS's internal world model. The REM aggregates multi-domain data including:

• Spectrum occupancy measurements from environmental sensing capability (ESC) sensors.
• Terrain and propagation models to calculate path loss.
• Registered transmitter locations and their technical parameters. The SAS queries the REM to perform propagation modeling and determine exclusion zones, ensuring that a new GAA or PAL assignment will not create harmful interference at any protected incumbent receiver.

A dedicated sensor network that forms the SAS's eyes and ears. ESCs are deployed along coastlines to detect federal incumbent radar systems operating in the 3.5 GHz band. When an ESC detects a radar pulse, it immediately notifies the SAS, which must then reconfigure or suspend secondary transmissions in the affected geographic area within a strict time limit. This sensor-driven dynamic protection zone is the mechanism that allows the Navy to maintain operational security while sharing spectrum with commercial users.

The proactive intelligence layer that transforms the SAS from a reactive coordinator into a predictive orchestrator. By applying recurrent neural networks and transformer models to historical spectrum usage data, the SAS can forecast future channel availability. This enables proactive channel assignment, where a device is pre-authorized on a frequency predicted to be vacant, dramatically reducing the latency of spectrum handoffs and improving quality of service for mobile secondary users.