Spectrum Sensing

The most fundamental non-coherent sensing method. It calculates the total received signal energy within a target band over a finite observation interval and compares it to a pre-calculated threshold. This technique requires no prior knowledge of the primary user's signal characteristics.

• Advantage: Low computational complexity and fast detection time.
• Critical Limitation: Highly susceptible to noise uncertainty. In low SNR environments, distinguishing a weak signal from background noise becomes statistically unreliable.
• Vulnerability: Cannot differentiate between a primary user, an interferer, or a burst of noise, leading to a high probability of false alarm.

The optimal detection method when the primary user's transmitted waveform is known a priori. It operates by correlating the received signal with a local copy of the known pilot, preamble, or spreading code.

• Mechanism: Maximizes the signal-to-noise ratio (SNR) at the decision point, enabling reliable detection even at very low power levels.
• Trade-off: Requires perfect synchronization and demodulation of the target signal, demanding a dedicated receiver for every potential primary user type.
• Use Case: Ideal for detecting known broadcast signals like ATSC pilots or LTE cell-specific reference signals where the waveform structure is standardized.

Exploits the inherent periodicity in modulated signals caused by carrier frequencies, symbol rates, and cyclic prefixes. Unlike stationary noise, modulated signals exhibit spectral correlation at specific cycle frequencies.

• Robustness: Highly immune to noise uncertainty because noise is generally stationary (no cyclic features), while modulated signals are cyclostationary.
• Capability: Can distinguish between different modulation types (e.g., BPSK vs. QPSK) based on their unique cyclic signatures, enabling signal classification alongside detection.
• Cost: Significantly higher computational complexity than energy detection due to the calculation of the Spectral Correlation Function (SCF).

A blind detection method that analyzes the eigenvalues of the sample covariance matrix of the received signal. It leverages the fact that a correlated primary user signal alters the eigenvalue distribution compared to uncorrelated white noise.

• Algorithms: Includes the Maximum-Minimum Eigenvalue (MME) test and the Energy with Minimum Eigenvalue (EME) test.
• Key Benefit: Overcomes the noise uncertainty problem without requiring any prior knowledge of the signal, channel, or noise power.
• Application: Effective for detecting spread-spectrum signals or OFDM transmissions where the cyclic prefix induces signal correlation, even when the exact waveform is unknown.

Enables the direct digitization and detection of multi-GHz wideband spectrum using sampling rates far below the Nyquist limit. It relies on the principle that the spectrum is inherently sparse—only a small fraction of frequencies are occupied at any instant.

• Architecture: Uses an Analog-to-Information Converter (AIC) to acquire compressed measurements, followed by L1-norm minimization algorithms (e.g., Basis Pursuit) to reconstruct the spectral support.
• Advantage: Drastically reduces the hardware cost, power consumption, and data throughput requirements of wideband receivers.
• Challenge: Reconstruction fidelity degrades if the sparsity assumption is violated by a sudden surge in spectral activity.

Combats the hidden node problem caused by multipath fading and shadowing by fusing sensing data from multiple geographically distributed nodes. Spatial diversity dramatically improves the reliability of detection.

• Hard Fusion: Nodes transmit local binary decisions (1-bit) to a fusion center, which applies a logical rule like OR, AND, or K-out-of-N.
• Soft Fusion: Nodes transmit their full test statistics or likelihood ratios, allowing the fusion center to perform optimal combining, often using Likelihood Ratio Test (LRT) weighting.
• Trade-off: Soft fusion provides superior performance but consumes significantly more control channel bandwidth.

The foundational non-coherent sensing method that measures the total power within a target band. It compares the accumulated energy against a dynamically calculated threshold to declare a channel occupied or vacant. Key characteristics:

• Complexity: Extremely low computational overhead, requiring no prior knowledge of the primary user signal.
• Vulnerability: Highly susceptible to the noise uncertainty problem, where fluctuating noise floors severely degrade detection reliability at low Signal-to-Noise Ratios (SNR).
• Application: Best suited for initial coarse sensing in high-SNR environments before triggering more sophisticated cyclostationary or matched filter detectors.

A robust detection method that exploits the inherent periodicity in modulated signals. Unlike stationary noise, communication signals exhibit statistical properties—such as mean and autocorrelation—that vary periodically with time. Operational advantages:

• Noise immunity: Functions reliably well below the noise floor where energy detectors fail, as it searches for specific spectral correlation patterns unique to the modulation scheme.
• Signal classification: Can simultaneously detect and classify the signal type (e.g., BPSK vs. QAM) by analyzing the cycle frequencies present in the Spectral Correlation Function (SCF).
• Computational cost: Requires high-resolution FFT operations and significant processing power, making it a trade-off between accuracy and hardware complexity.

A distributed architecture that mitigates the hidden node problem caused by multipath fading and shadowing. Multiple spatially separated sensing nodes share their local observations with a fusion center to make a global decision. Fusion strategies include:

• Hard combining: Nodes transmit binary 1-bit decisions (occupied/idle) to the center, which applies logical AND, OR, or K-out-of-N voting rules.
• Soft combining: Nodes transmit their raw energy levels or likelihood ratios, allowing the fusion center to perform optimal weighted combining for maximum sensitivity.
• Trade-off: Dramatically improves detection probability but introduces overhead for a dedicated control channel and requires precise synchronization among cooperating nodes.

The theoretically optimal detection method when the secondary user possesses perfect a priori knowledge of the primary user's transmitted waveform. It maximizes the output SNR by correlating the received signal with a time-reversed replica of the known pilot or preamble. Critical constraints:

• Coherent detection: Requires precise timing and carrier synchronization to align the filter template with the incoming signal.
• Knowledge requirement: Demands a dedicated receiver for every potential primary signal format, making it impractical for heterogeneous spectrum environments.
• Performance: Achieves the lowest sensing time for a given probability of false alarm, making it the gold standard benchmark against which blind sensing methods are measured.

A sub-Nyquist sampling paradigm that enables direct digitization of multi-GHz spectrum spans without prohibitive ADC hardware costs. It leverages the inherent sparsity of spectrum usage—where only a fraction of channels are active simultaneously. Core mechanism:

• Sparse recovery: Uses optimization algorithms like Basis Pursuit or Orthogonal Matching Pursuit to reconstruct the full spectrum from heavily undersampled measurements.
• Hardware enabler: The Modulated Wideband Converter (MWC) is a practical implementation that mixes the input with pseudo-random sequences before low-rate sampling.
• Advantage: Eliminates the need for slow, power-hungry swept-tuned scanners, enabling real-time situational awareness across the entire tactical communications band.

A blind sensing technique that distinguishes a correlated primary user signal from uncorrelated white noise by analyzing the statistical structure of the received sample covariance matrix. Methodology:

• Test statistics: Derives ratios from the matrix eigenvalues (e.g., Maximum-to-Minimum Eigenvalue, MME) or off-diagonal energy. The presence of a signal introduces correlation that inflates these metrics.
• Noise immunity: Completely bypasses the noise uncertainty problem because it does not rely on an absolute noise power estimate for threshold setting.
• Multi-antenna synergy: Performance scales naturally with the number of receiver antennas, as the dimensionality of the covariance matrix increases, providing robust detection even with severely imbalanced IQ data.