Ring Ratio

The ring ratio (often denoted as ρ or γ) is the quotient of the radii of concentric amplitude rings in an APSK or circular QAM constellation. It is mathematically expressed as ρ = R_outer / R_inner, where R_outer is the radius of the ring with the larger amplitude and R_inner is the radius of the inner ring. This dimensionless parameter defines the relative spacing between energy levels and is a critical geometric property that must be estimated or known a priori for accurate minimum distance decoding and modulation format identification.

The ring ratio directly governs the peak-to-average power ratio (PAPR) of the transmitted signal. A ratio close to 1.0 compresses the rings together, minimizing PAPR but reducing the Euclidean distance between points on different rings, which degrades error performance in additive white Gaussian noise. A larger ratio increases the inter-ring distance, improving noise immunity for amplitude-dependent bits, but demands more linearity from the high-power amplifier (HPA). Optimizing this ratio is a fundamental trade-off in geometric shaping for non-linear satellite channels.

In automatic modulation classification (AMC), the ring ratio serves as a robust, hand-crafted feature for hierarchical classifiers. By applying k-means clustering or centroid estimation to the magnitudes of received IQ samples, a system can estimate the number of rings and their relative radii. This estimated ratio is then compared against a library of theoretical templates for known formats (e.g., 16-APSK vs. 32-APSK). This method is particularly effective because the ratio is invariant to carrier phase offset and, to a first order, phase ambiguity.

Accurate blind estimation of the ring ratio is challenged by IQ imbalance and non-linear distortion. IQ imbalance causes the circular rings to warp into ellipses, making a simple radial average unreliable. Non-linear amplification compresses the outer ring more than the inner ring, altering the effective ratio at the receiver. Robust estimation requires pre-compensation using blind equalization algorithms like the Constant Modulus Algorithm (CMA) or a modified multi-modulus algorithm to restore the circular ring structure before computing the ratio.

The DVB-S2 and DVB-S2X satellite broadcasting standards define specific ring ratios for their APSK constellations to balance spectral efficiency and HPA non-linearity tolerance. For example:

• 16-APSK: The canonical ring ratio (R2/R1) is often 2.57 or 3.15 depending on the code rate.
• 32-APSK: Uses three rings with defined ratios (e.g., R2/R1 ≈ 2.64, R3/R1 ≈ 4.64).
• 64-APSK: Employs four rings with carefully optimized ratios. These pre-defined values serve as a strong prior for template matching classifiers.

While both APSK and circular QAM use concentric rings, the ring ratio philosophy differs. APSK typically uses a small number of rings (2-4) with many phase states per ring, and the ring ratio is a primary design parameter. Circular QAM often uses many rings with fewer points per ring to approximate a Gaussian distribution for probabilistic shaping. In circular QAM, the ring ratios are often uniform or follow a specific distribution to maximize mutual information, rather than being a single defining parameter.

The modulation family where ring ratio is the defining geometric parameter. APSK arranges symbols on multiple concentric rings of uniform amplitude, with varying phase states per ring.

• Ring Ratio (ρ): Defined as the radius of the outer ring (R₂) divided by the radius of the inner ring (R₁)
• Motivation: Reduces peak-to-average power ratio (PAPR) compared to square QAM, critical for non-linear satellite transponders
• Common Formats: 16-APSK (4+12), 32-APSK (4+12+16), 64-APSK (4+12+20+28)
• Standardization: DVB-S2/S2X specifies optimal ring ratios for each code rate to maximize mutual information

The primary computational method for blind ring ratio estimation. By clustering received IQ samples and calculating the geometric center of each amplitude ring, the ring ratio can be derived without prior knowledge of the modulation format.

• Process: K-means or Gaussian Mixture Models partition samples into amplitude clusters
• Radius Calculation: Euclidean distance from origin to each cluster centroid
• Ratio Derivation: ρ̂ = R̂_outer / R̂_inner
• Application: Enables autonomous modulation format identification in cognitive radio and spectrum monitoring systems

A quantitative metric directly impacted by ring ratio estimation accuracy. EVM measures the Euclidean distance between the ideal reference constellation point and the actual received symbol.

• Ring Ratio Dependency: Incorrect ring ratio assumptions cause systematic EVM degradation on outer rings
• Calculation: EVM_RMS = sqrt(Σ|S_measured - S_ideal|² / N) normalized to constellation peak amplitude
• Thresholds: DVB-S2 requires EVM < 7.5% for 32-APSK with low code rates
• Diagnostic Value: EVM vs. ring index plots reveal if ring ratio misestimation is the dominant impairment

A blind equalization technique that can distort the ring ratio of APSK constellations if not properly modified. Standard CMA penalizes deviations from a single constant modulus, collapsing multiple rings toward a single amplitude.

• Multi-Modulus CMA (MMCA): Extension that uses separate modulus targets for each ring, preserving the ring ratio
• Radius-Directed Equalization: Explicitly incorporates estimated ring radii into the cost function
• Convergence: Requires initial ring ratio estimate or hierarchical switching from CMA to decision-directed mode
• Application: Critical for satellite links where training sequences are unavailable

The optimization framework that treats ring ratio as a continuous design variable rather than a fixed parameter. Geometric shaping moves constellation points in the IQ plane to maximize channel capacity.

• Optimization Objective: Maximize generalized mutual information (GMI) for a specific SNR and channel model
• Ring Ratio Optimization: Non-uniform ring spacing often outperforms uniform ratios in non-linear channels
• Iterative Algorithms: Pairwise optimization or gradient descent on point locations
• Result: Shaped constellations with irregular ring ratios that approach Shannon capacity within 0.1 dB

A modulation classification method that compares a reconstructed received constellation against a library of ideal reference constellations with known ring ratios.

• Correlation Process: Cross-correlate received IQ histogram with each template after scale and rotation compensation
• Ring Ratio as Discriminant: The ratio uniquely identifies APSK orders (e.g., 16-APSK vs 32-APSK have different ring counts and ratios)
• Similarity Metrics: Normalized cross-correlation coefficient or Bhattacharyya distance between distributions
• Robustness: Performs well at moderate SNR (>10 dB) where ring structure is visually discernible