Missed Detection Probability

The most fundamental driver of missed detection. When the primary user's signal power at the sensing receiver is below the noise floor, distinguishing signal from noise becomes statistically unreliable.

• Hidden Node Problem: Physical obstructions or deep fades can attenuate the primary signal by 20-30 dB, rendering it undetectable even at close range.
• Noise Uncertainty: In practice, the noise floor is not perfectly known. A 1 dB error in noise estimation can cause a sharp increase in missed detections at low SNR.
• Mitigation: Longer sensing durations and cooperative sensing architectures can partially compensate, but the fundamental energy detection limit persists.

The sensing-throughput tradeoff imposes a hard limit on observation time. Cognitive radios must vacate the channel quickly to sense, and longer sensing improves detection but reduces transmission capacity.

• Finite Sample Effects: With limited samples, the sample covariance matrix is a noisy estimate of the true covariance, degrading eigenvalue-based detection methods.
• Coherent vs. Non-Coherent Integration: Coherent integration provides higher processing gain but requires accurate synchronization. Non-coherent integration is more robust but less efficient.
• IEEE 802.22 Standard: Specifies a maximum sensing time of ≤ 2 seconds for TV band detection, creating a strict upper bound on integration gain.

Non-ideal analog front-end components introduce distortions that can mask or corrupt the primary signal, increasing missed detection probability even at favorable SNR.

• IQ Imbalance: Gain and phase mismatches between in-phase and quadrature branches create ghost signals that confuse cyclostationary feature detectors.
• Phase Noise: Local oscillator jitter spreads the signal spectrum, reducing the effective SNR for narrowband detection algorithms.
• Nonlinearities: Amplifier saturation and intermodulation products can generate spurious signals that either mask the primary user or trigger false alarms, forcing the use of conservative detection thresholds that increase missed detections.
• ADC Quantization: Low-resolution analog-to-digital converters introduce quantization noise that sets a floor on achievable sensitivity.

Wireless propagation effects cause the received signal power to fluctuate randomly over time, frequency, and space, creating deep fades where detection becomes impossible.

• Rayleigh Fading: In non-line-of-sight environments, signal amplitude follows a Rayleigh distribution, with frequent deep nulls of 30-40 dB below the mean.
• Shadowing: Large-scale obstructions create log-normal shadowing with standard deviations of 6-12 dB in urban environments.
• Frequency-Selective Fading: Wideband signals experience different fading across subcarriers. A narrowband sensing algorithm may sample a faded portion of the spectrum and miss the primary user entirely.
• Mitigation: Cooperative sensing with spatial diversity is the primary countermeasure, as the probability of simultaneous deep fades at multiple receivers is low.

The choice of detection threshold directly controls the tradeoff between false alarm probability (Pfa) and missed detection probability (Pmd). Setting the threshold is a statistical decision under uncertainty.

• Constant False Alarm Rate (CFAR): Adaptive thresholding techniques maintain a fixed Pfa by estimating the local noise floor, but estimation errors propagate directly to Pmd.
• Neyman-Pearson Criterion: The optimal approach maximizes detection probability for a given Pfa constraint, but requires accurate knowledge of the signal distribution under both hypotheses.
• Regulatory Requirements: IEEE 802.22 mandates a detection probability of ≥ 90% for TV signals at -116 dBm, forcing threshold settings that may increase false alarms but prioritize primary user protection.
• Model Mismatch: If the assumed noise distribution is incorrect, the actual Pmd can deviate significantly from the designed value.

The structure and behavior of the primary user's waveform directly impact detectability. Signals with low peak-to-average power ratio (PAPR) or irregular transmission patterns are inherently harder to detect.

• Spread Spectrum Signals: Direct-sequence spread spectrum (DSSS) and frequency-hopping signals intentionally spread energy below the noise floor, making energy detection ineffective. Cyclostationary feature detection is required.
• Intermittent Transmission: Primary users with bursty traffic patterns may be silent during the sensing window, leading to a false sense of vacancy and subsequent collision when transmission resumes.
• Unknown Modulation: If the sensing algorithm is matched to a specific modulation type, mismatched primary signals will suffer degraded detection performance.
• Dynamic Power Control: Primary users that adapt their transmit power based on their own link budget create a moving target for threshold-based detectors.

The probability that a spectrum sensing algorithm incorrectly declares a frequency band as occupied when it is actually vacant. This Type I error directly reduces spectral efficiency by causing secondary users to miss transmission opportunities.

• Tradeoff: Inversely related to missed detection probability via the receiver operating characteristic (ROC) curve
• Impact: High false alarm rates waste spectrum holes and reduce cognitive radio throughput
• Threshold tuning: Lowering the detection threshold reduces missed detections but increases false alarms
• Neyman-Pearson criterion: Classical approach that fixes false alarm rate and minimizes missed detection probability

A graphical plot that illustrates the diagnostic ability of a spectrum sensing detector by mapping detection probability against false alarm rate as the decision threshold varies. The ROC curve is the fundamental tool for evaluating and comparing sensing algorithms.

• Area Under Curve (AUC): Single scalar metric summarizing detector performance; AUC = 1.0 represents perfect classification
• Operating point selection: Engineers choose a specific point on the ROC curve based on regulatory requirements for primary user protection
• SNR dependence: ROC curves degrade significantly at low signal-to-noise ratios, directly increasing missed detection probability
• Diversity gain: Cooperative sensing and multiple antennas shift the ROC curve upward and leftward

A sensing vulnerability where a secondary user is physically obstructed from detecting a primary transmitter due to terrain, buildings, or fading, causing unintentional harmful interference. This phenomenon directly increases missed detection probability regardless of sensing algorithm quality.

• Shadowing: Large-scale fading behind obstacles attenuates the primary signal below the detection threshold
• Multipath fading: Constructive and destructive interference creates deep fades that mask primary transmissions
• Mitigation: Cooperative spectrum sensing with geographically distributed nodes is the primary countermeasure
• Exposed node counterpart: The opposite problem where a secondary user unnecessarily defers transmission due to detecting a distant primary that would not experience interference

A collaborative detection framework where multiple secondary users share local sensing observations with a fusion center to improve primary user detection reliability. This approach directly combats missed detections caused by hidden nodes and deep fading.

• Hard combining: Nodes transmit binary decisions (occupied/vacant); fusion center applies AND, OR, or K-out-of-N rules
• Soft combining: Nodes transmit raw energy measurements or likelihood ratios for optimal weighted fusion
• Diversity gain: Spatial diversity from distributed sensors dramatically reduces the probability of simultaneous missed detection
• Overhead tradeoff: Reporting channel bandwidth consumption must be balanced against detection performance gains

A statistical hypothesis testing framework that constrains the false alarm probability to a maximum acceptable level while minimizing the missed detection probability. This is the theoretical foundation for designing optimal spectrum sensing detectors under regulatory constraints.

• Likelihood ratio test: The optimal detector compares the ratio of signal-present to signal-absent probability densities against a threshold
• Constant False Alarm Rate (CFAR): Adaptive threshold techniques that maintain fixed false alarm rates despite changing noise floors
• Regulatory mapping: The IEEE 802.22 standard requires detection probability ≥ 90% at SNR levels as low as -20 dB for TV band devices
• Energy detection limitation: The simple energy detector cannot simultaneously optimize both error probabilities without knowing the signal structure

The process by which a secondary user vacates its current channel upon detecting a returning primary user and seamlessly transitions to another available idle channel. Missed detection probability directly determines the latency and success rate of this critical mobility function.

• Reactive handoff: Triggered only after primary user detection; vulnerable to missed detections causing interference
• Proactive handoff: Predictive channel selection based on historical occupancy patterns to reserve backup channels in advance
• Handoff latency: Target should remain below the maximum allowed interference duration specified by regulators
• Connection preservation: Higher-layer protocols must maintain session continuity during physical layer channel switching