Sensing-Throughput Tradeoff

The fundamental structural constraint. A cognitive radio frame is divided into a sensing slot and a data transmission slot. Increasing the sensing duration improves detection reliability but directly reduces the time available for throughput.

• Periodic Sensing Frame: The total frame duration is fixed, creating a zero-sum allocation problem.
• Optimal Sensing Time: Exists at the point where the marginal gain in detection probability no longer justifies the loss in transmission opportunity.
• MAC Layer Design: The frame structure must be designed to accommodate the worst-case sensing time required to meet regulatory detection thresholds, such as IEEE 802.22's requirement for TV band detection.

Regulatory bodies mandate a minimum probability of detection to protect incumbent primary users. This constraint acts as a hard floor that the sensing algorithm must meet, regardless of its impact on throughput.

• Regulatory Threshold: The IEEE 802.22 standard requires a 90% probability of detection for TV signals at -116 dBm.
• Missed Detection Penalty: A missed detection causes harmful interference to a licensed user, which is an unacceptable failure mode.
• Constraint-Driven Design: The sensing algorithm must be configured to satisfy this constraint first; throughput is maximized only within the remaining design space.

A false alarm occurs when the sensor declares a channel occupied when it is actually vacant. This directly wastes a transmission opportunity and reduces throughput.

• Opportunity Cost: Every false alarm represents a spectrum hole that was available but not utilized.
• Tradeoff Coupling: For a fixed sensing duration, lowering the false alarm probability typically requires raising the detection threshold, which increases the missed detection probability.
• Throughput Sensitivity: In sparse spectrum environments with many idle channels, the false alarm rate is the dominant factor limiting secondary throughput.

The ROC curve maps the relationship between probability of detection and probability of false alarm for a given sensing algorithm and SNR. The operating point on this curve is the primary design choice.

• Operating Point Selection: The optimal point maximizes throughput while satisfying the detection constraint.
• Algorithm Comparison: Different sensing methods—energy detection, cyclostationary detection, matched filtering—exhibit different ROC characteristics, with more sophisticated methods offering better tradeoff curves at the cost of complexity.
• SNR Dependence: The ROC curve shifts dramatically with received signal strength, meaning the optimal operating point is environment-dependent.

The received Signal-to-Noise Ratio at the cognitive radio is unknown and time-varying due to fading, shadowing, and mobility. This uncertainty forces conservative sensing configurations.

• SNR Wall: Below a critical SNR threshold, reliable detection becomes theoretically impossible regardless of sensing duration, due to noise uncertainty.
• Fading Margins: The sensing algorithm must be designed for worst-case channel conditions, sacrificing throughput during favorable conditions.
• Adaptive Sensing: Advanced systems dynamically adjust sensing time based on real-time SNR estimates to optimize the instantaneous tradeoff.

In cooperative spectrum sensing, multiple nodes share observations to mitigate the hidden node problem. However, this introduces additional overhead that impacts the effective tradeoff.

• Reporting Delay: Time spent transmitting sensing reports to a fusion center reduces the time available for data transmission.
• Fusion Rule Impact: Hard decision fusion minimizes overhead but loses information; soft decision fusion preserves information but consumes more bandwidth.
• Node Selection: Optimizing the number of cooperating nodes balances diversity gain against coordination overhead, directly affecting the net throughput.

The conditional probability that a sensing algorithm correctly declares a frequency band as occupied when a primary user signal is truly present. This is the most critical performance metric in cognitive radio design.

• Mathematical definition: P_d = P(decision = occupied | signal present)
• Regulatory requirement: IEEE 802.22 mandates P_d ≥ 90% for TV white space devices
• Tradeoff impact: Increasing P_d requires longer sensing time or more complex algorithms, directly reducing available transmission time
• Complementary metric: Missed detection probability = 1 - P_d represents the probability of causing harmful interference

The conditional probability that a sensing algorithm incorrectly declares a vacant band as occupied, causing the cognitive radio to miss a transmission opportunity.

• Mathematical definition: P_fa = P(decision = occupied | signal absent)
• Throughput impact: Every false alarm directly wastes a spectrum hole that could have carried data
• Tradeoff relationship: P_fa and P_d are coupled—lowering the detection threshold increases both
• Neyman-Pearson criterion: The optimal detector maximizes P_d subject to a constraint on P_fa, formalizing the sensing-throughput tradeoff as a constrained optimization problem

A graphical plot that visualizes the sensing-throughput tradeoff by showing the relationship between probability of detection and false alarm probability as the decision threshold varies.

• ROC curve: Plots P_d (y-axis) against P_fa (x-axis) for all possible threshold values
• Area Under Curve (AUC): A single scalar metric summarizing detector performance; AUC = 1.0 represents perfect classification
• Practical use: Engineers select an operating point on the ROC curve that balances regulatory P_d requirements against throughput-maximizing P_fa targets
• SNR dependence: Higher signal-to-noise ratios shift the ROC curve toward the upper-left corner, enabling better tradeoff outcomes

A distributed architecture where multiple cognitive radios share local sensing observations to mitigate the hidden node problem and improve overall detection reliability, fundamentally altering the sensing-throughput tradeoff.

• Spatial diversity: Multiple geographically separated sensors overcome shadowing and multipath fading that blind individual nodes
• Sensing time reduction: Cooperation enables shorter per-node sensing durations while maintaining target P_d, directly increasing throughput
• Fusion strategies: Hard decision fusion (AND, OR, K-out-of-N rules) trades bandwidth efficiency for detection performance; soft decision fusion preserves more information
• Overhead cost: Reporting local decisions to a fusion center consumes bandwidth that must be accounted for in the net throughput calculation

The inherent fluctuation in ambient noise power that fundamentally limits the performance of energy detectors, creating an SNR wall below which reliable detection is impossible regardless of sensing duration.

• Source: Thermal noise variations, amplifier gain fluctuations, and environmental interference cause noise power to deviate from nominal values
• SNR Wall: The theoretical minimum SNR threshold below which a non-coherent detector cannot achieve arbitrary P_d and P_fa simultaneously, no matter how long it senses
• Tradeoff implication: In noise-limited regimes, increasing sensing time yields diminishing returns—the throughput sacrifice cannot be compensated by improved detection
• Mitigation: Cyclostationary feature detection and eigenvalue-based methods are robust to noise uncertainty but require higher computational complexity

An AI-driven approach where an agent learns an optimal sensing policy through trial-and-error interaction with the environment, dynamically adapting sensing parameters to maximize throughput while minimizing interference.

• State space: Includes channel occupancy history, sensed energy levels, and throughput statistics
• Action space: The agent selects sensing duration, detection threshold, and transmission power for each frame
• Reward function: Positively rewards successful data transmission and penalizes missed detections that cause interference
• Adaptive advantage: Unlike static frame designs, DRL policies learn to shorten sensing time in historically vacant channels and extend it in crowded or uncertain bands, achieving a superior dynamic tradeoff