On High-Order Capacity Statistics of Spectrum Aggregation Systems over $κ$-$μ$ and $κ$-$μ$ shadowed Fading Channels

Jiayi Zhang, Xiaoyu Chen, Kostas P. Peppas, Xu Li, Ying Liu

Introduction

As the demand for mobile data continues to grow exponentially toward 5G and beyond, the scarcity of contiguous radio spectrum has become a significant bottleneck. Spectrum Aggregation (SA) has emerged as a critical solution, allowing network operators to "stitch together" non-adjacent frequency bands into a larger virtual pipe. While previous research has extensively explored the ergodic (average) capacity of such systems, average performance metrics often fail to capture the reliability and stability required for modern high-speed wireless links.

System model for spectrum aggregation over multiple frequency bands under generalized fading conditions.

This paper addresses this gap by providing a rigorous mathematical framework to analyze the Higher-Order Statistics (HOS) of channel capacity. By examining variance, skewness, and kurtosis, the authors offer a more nuanced understanding of "jitter" and the probability of service drops in realistic propagation environments.

System Model and Fading Environments

The study focuses on SA systems operating over generalized fading channels, specifically the $\kappa$-$\mu$ and $\kappa$-$\mu$ shadowed models. These models are highly flexible:

  • $\kappa$-$\mu$ Distribution: Represents multipath clusters with a dominant line-of-sight (LoS) component.
  • $\kappa$-$\mu$ Shadowed Distribution: Accounts for fluctuations in the LoS signal due to obstacles, providing a more accurate representation of 5G urban environments.

The aggregation of $M$ bands, each with its own signal-to-noise ratio (SNR) $\gamma_i$, forms the basis of the analysis.

Analytical Methodology

The authors derive exact closed-form expressions for the HOS of channel capacity using a purely analytical approach. The mathematical framework leverages:

  • Probability Density Functions (PDFs) and Moment Generating Functions (MGFs): To characterize the statistical behavior of the aggregated bands.
  • Advanced Special Functions: Solutions involve Meijer’s G-functions, Tricomi hypergeometric functions, and Stirling numbers of the first kind to handle complex integrals.
  • Numerical Efficiency: A 15-point Gauss-Chebyshev quadrature rule is implemented to ensure rapid computation, significantly outperforming traditional Monte-Carlo simulations in terms of speed without sacrificing precision.
  • Asymptotic Analysis: Simplified formulas for low-SNR and high-SNR regimes provide intuitive insights into system behavior at extreme signal levels.

Key Insights and Results

The analytical derivations were validated against Monte-Carlo simulations using $10^6$ random realizations, showing a perfect match across all tested SNR ranges.

Beyond the Average

The calculation of skewness and kurtosis reveals that capacity distributions are often asymmetric and "spiky." This implies that high average capacity does not guarantee reliability; the probability of the rate falling below a usable threshold remains a critical factor for engineers to manage.

Diversity Gain and Stability

A key finding is that increasing the number of aggregated bands ($M$) does more than just linearly increase capacity; it significantly reduces the Amount of Dispersion (AoD). This "diversity gain" means that more aggregated bands lead to a more stable and deterministic connection. For instance, a sample 3-band system at 9 dB SNR achieved a 90.81% reliability percentage.

Impact of Shadowing

At low SNR, the shadowing parameter ($m$) emerges as the dominant factor affecting reliability. This suggests that in power-constrained scenarios, improving line-of-sight conditions is more effective than simply increasing transmission power.

Conclusion

This research provides essential tools for the design and optimization of 5G and future 6G networks. By moving beyond ergodic capacity to higher-order statistics, engineers can better predict and improve the reliability of spectrum aggregation systems. Future work may extend this framework to Multiple-Input Multiple-Output (MIMO) systems and the development of dynamic channel selection algorithms to further enhance network performance.

Summary updated

BibTeX
@article{zhang2017high,
  author = {Jiayi Zhang and Xiaoyu Chen and Kostas P. Peppas and Xu Li and Ying Liu},
  title = {{On High-Order Capacity Statistics of Spectrum Aggregation Systems over $κ$-$μ$ and $κ$-$μ$ shadowed Fading Channels}},
  journal = {IEEE Trans. Communications},
  year = {2017},
  doi = {10.1109/tcomm.2016.2637926},
}
Figure