I/Q Gain Ratio

The I/Q gain ratio is defined as G_I / G_Q, where G_I is the amplitude gain of the in-phase path and G_Q is the amplitude gain of the quadrature path. In an ideal direct-conversion transmitter, this ratio is exactly 1.0 (0 dB). Any deviation indicates a gain imbalance.

• Ratio > 1: The I-channel has higher gain, stretching the constellation horizontally.
• Ratio < 1: The Q-channel has higher gain, stretching the constellation vertically.
• Unit: Often expressed in decibels (dB) as 20 * log10(G_I / G_Q).

A gain imbalance directly causes a rectangular distortion of the ideal constellation. For a 16-QAM signal, a gain ratio of 1.2 (1.58 dB) transforms the perfect square grid into a visibly stretched rectangle.

• Visual Signature: The outer symbol points no longer form a square; the aspect ratio changes.
• EVM Impact: The Error Vector Magnitude increases because the received symbols are displaced radially from their ideal reference targets.
• Detection: This error is easily visible on an I/Q constellation diagram as an asymmetry between the I and Q axes.

Gain mismatch originates in the analog baseband and RF front-end components. Primary sources include:

• DAC Mismatch: Slight differences in the full-scale output current or voltage reference between the I and Q digital-to-analog converters.
• Baseband Amplifier Tolerance: Variations in the gain-setting resistor values in the reconstruction filters and variable gain amplifiers (VGAs).
• Mixer Conversion Loss: Asymmetrical conversion loss in the I and Q channels of the quadrature modulator.
• PCB Trace Impedance: Minor impedance variations in the differential I and Q transmission lines.

The I/Q gain ratio is a highly stable, device-specific impairment that forms a cornerstone of physical layer authentication. Because it is determined by fixed manufacturing tolerances in passive and active components, it acts as an unclonable hardware identifier.

• Uniqueness: The exact gain error (e.g., 0.87 dB) is statistically unique per device.
• Stability: The ratio remains constant over short time frames under fixed temperature conditions.
• Feature Vector: Used alongside quadrature skew and DC offset to construct a multi-dimensional I/Q distortion signature for machine learning classifiers.

Blind estimation of the I/Q gain ratio is performed directly on the received signal without a known training sequence. A common statistical method leverages the property that ideal I and Q components are uncorrelated and have equal variance.

• Statistical Method: The gain ratio is estimated as sqrt( E[Q^2] / E[I^2] ), where E[] denotes the expected value.
• Test Equipment: Vector signal analyzers (VSAs) compute this automatically from the demodulated constellation.
• Precision: Modern algorithms can estimate the ratio with an accuracy of < 0.1 dB.

Gain imbalance is a linear impairment and can be perfectly corrected in the digital domain using an adaptive I/Q correction matrix before the DAC.

• Correction Matrix: A 2x2 matrix multiplies the I/Q samples to restore orthogonality and amplitude balance.
• Calibration: Factory calibration routines measure the native gain ratio and store the inverse correction coefficients in non-volatile memory.
• Trade-off: While correction cleans the transmitted signal for communication purposes, it can mask the unique fingerprint used for security, requiring the authenticator to model the pre-correction impairment.

The compound hardware impairment where the in-phase (I) and quadrature (Q) signal paths exhibit mismatched amplitude (gain ratio ≠ 1) or phase deviation (quadrature skew ≠ 0°). This creates a deterministic, device-specific distortion pattern in the constellation diagram that serves as a primary fingerprinting feature. The gain ratio component scales symbols along one axis, while phase error introduces cross-talk between I and Q channels.

A direct visual manifestation of I/Q gain ratio imbalance where the constellation diagram compresses or expands along one axis. Key characteristics:

• Gain ratio > 1: I-channel amplitude exceeds Q-channel, stretching symbols horizontally
• Gain ratio < 1: Q-channel amplitude dominates, stretching symbols vertically
• Impact: Alters the amplitude ratio of symbols, degrading modulation fidelity and creating a measurable, unique signature

The deviation of the phase difference between I and Q local oscillator signals from the ideal 90 degrees. While distinct from gain ratio, quadrature skew interacts with amplitude imbalance to produce constellation warping—transforming a square grid into a parallelogram. The combined gain-phase impairment matrix is highly individual to each transmitter's analog front-end components.

A quantitative measure of how a nominally circular constellation point cluster stretches into an ellipse due to combined gain and phase imbalance. The ellipticity ratio directly correlates with the I/Q gain ratio, while the tilt angle of the ellipse's major axis reveals the phase imbalance. These geometric parameters form robust, channel-invariant features for machine learning-based emitter identification.

Digital signal processing techniques that dynamically estimate and compensate for time-varying I/Q imbalance. Methods include:

• Blind estimation algorithms that derive correction coefficients from received signal statistics
• Pilot-based calibration using known training sequences
• Feedback loops that continuously track gain ratio drift due to temperature and aging Effective correction is critical for distinguishing intentional modulation from hardware fingerprint artifacts.

The unique, repeatable pattern of constellation deformation caused by a specific transmitter's combination of gain ratio, quadrature skew, and DC offset. This multi-parameter profile remains stable under fixed conditions and serves as an unclonable physical-layer identifier. The signature's uniqueness across devices and stability over time are the foundational requirements for RF fingerprinting authentication systems.