Spectrum Sensing

The most common and computationally simple sensing technique that measures the total received signal energy within a target frequency band over a defined observation interval. The detector compares the accumulated energy against a pre-calculated threshold derived from the estimated noise floor. Key characteristics:

• Does not require any prior knowledge of the primary user's signal structure
• Performs poorly under low Signal-to-Noise Ratio (SNR) conditions, a phenomenon known as the SNR Wall
• Highly susceptible to noise uncertainty, where even a 1 dB error in noise floor estimation can cause a catastrophic drop in detection probability
• Optimal for detecting any zero-mean constellation signal when noise power is perfectly known

A coherent detection method that correlates the received signal with a known replica of the primary user's transmitted waveform to maximize the output SNR. This technique requires a priori knowledge of the signal's pilot tone, preamble, or spreading code. Performance attributes:

• Achieves the optimal detection performance in additive white Gaussian noise
• Requires only O(1/SNR) samples to achieve a target probability of error, making it extremely fast
• Demands perfect synchronization (timing and carrier) and a dedicated receiver for each primary user signal type
• Considered impractical for general-purpose spectrum sensing due to its high signaling overhead and inflexibility

Exploits the periodic statistical properties inherent in modulated signals to distinguish them from stationary noise. Most communication signals exhibit cyclostationarity due to sine wave carriers, pulse trains, or cyclic prefixes, generating spectral correlation at specific cyclic frequencies. Advantages:

• Robust against noise uncertainty because noise is generally stationary (WSS) and exhibits no cyclic features
• Can differentiate between primary user types by identifying their unique cyclic signatures
• Significantly higher computational complexity than energy detection, requiring calculation of the Spectral Correlation Function (SCF)
• Effective at very low SNRs where energy detectors fail entirely

A blind sensing method that analyzes the eigenvalues of the sample covariance matrix constructed from signals received across multiple antennas or oversampled time samples. Core mechanisms:

• Maximum-Minimum Eigenvalue (MME): Computes the ratio of the largest to smallest eigenvalue; a ratio significantly greater than 1 indicates the presence of a correlated signal
• Energy with Minimum Eigenvalue (EME): Uses the smallest eigenvalue as a robust estimate of the noise variance, bypassing explicit noise estimation
• Completely immune to noise uncertainty because it derives the decision threshold from the signal statistics themselves
• Requires multiple receiver antennas or oversampling, increasing hardware complexity

A dedicated detection technique applicable when the primary user transmits a known pattern such as a preamble, midamble, or pilot sequence. The receiver performs a cross-correlation between the received signal and the known pattern. Operational details:

• Exploits the auto-correlation properties of the known sequence (e.g., Zadoff-Chu sequences in LTE)
• Provides highly reliable detection with very short sensing durations
• Commonly used in systems like Wi-Fi (802.11) where the Short Training Field (STF) preamble is standardized
• Limited applicability as it requires the primary network's frame structure to be known and static

An emerging class of sensing techniques that train supervised or unsupervised models to classify spectrum occupancy directly from raw IQ samples or extracted features. Common architectures:

• Convolutional Neural Networks (CNNs) for learning hierarchical features from time-frequency representations like spectrograms
• Autoencoders for unsupervised anomaly detection, flagging signals that deviate from the learned noise distribution
• Support Vector Machines (SVMs) for binary occupancy classification using hand-crafted features like energy, kurtosis, and cyclostationary signatures
• Excels in complex, dynamic environments where traditional model-based detectors fail due to unknown interference patterns

A distributed detection architecture where multiple spatially separated cognitive radios share local sensing observations with a fusion center. This approach overcomes the hidden node problem, where a single sensor might miss a primary transmitter due to shadowing or multipath fading. Fusion rules like AND, OR, or Majority Logic combine individual hard decisions, while soft combining transmits full energy levels for weighted aggregation, dramatically improving detection probability in fading environments.

A denial-of-service threat where a malicious actor mimics the signal characteristics of a licensed primary user to illegitimately reserve spectrum. The attacker replicates known features like pilot tones, cyclostationary signatures, or specific modulation formats. Defenses include RF fingerprinting to identify unique hardware imperfections and location verification using received signal strength (RSS) or angle-of-arrival to confirm the transmitter's claimed position matches the protected contour.

A fundamental sensing failure mode where a cognitive radio is shadowed from a primary transmitter by a physical obstruction like a building or hill. The secondary user incorrectly concludes the channel is vacant and transmits, causing harmful interference to a nearby primary receiver that is within its transmission range. Cooperative sensing architectures and higher detection sensitivity margins are the primary mitigations for this spatial uncertainty.

The process by which a secondary user vacates its current frequency channel upon detecting a returning primary user and seamlessly transitions to an alternative available channel. Key performance metrics include handoff latency (target: < 100 ms for real-time applications), probability of forced termination, and spectrum mobility prediction accuracy. Proactive handoff strategies pre-select backup channels to minimize session disruption.

A regulatory metric defined by the FCC that establishes the maximum tolerable interference level at a primary receiver. It sets an upper bound on the cumulative emissions secondary users may introduce into a licensed band. The model treats interference as a temperature-like field, where the goal is to keep the aggregate noise-plus-interference floor below a threshold that would degrade the primary receiver's service. This enables underlay spectrum sharing architectures.

A channel access mechanism requiring a transmitter to perform a Clear Channel Assessment (CCA) and verify the absence of other transmissions before initiating its own. Widely used in unlicensed spectrum sharing protocols like Wi-Fi (CSMA/CA) and LTE-LAA. The device samples the channel energy for a defined duration; if energy exceeds a threshold, it defers transmission and initiates a random backoff period to avoid collisions.