Outage Exponent: A Unified Performance Metric for Parallel Fading Channels

Bo Bai, Wei Chen, Khaled Ben Letaief, Zhigang Cao

Abstract

The parallel fading channel, which consists of finite number of subchannels, is very important, because it can be used to formulate many practical communication systems. The outage probability, on the other hand, is widely used to analyze the relationship among the communication efficiency, reliability, SNR, and channel fading. To the best of our knowledge, the previous works only studied the asymptotic outage performance of the parallel fading channel which are only valid for a large number of subchannels or high SNRs. In this paper, a unified performance metric, which we shall refer to as the outage exponent, will be proposed. Our approach is mainly based on the large deviations theory and the Meijer's G-function. It is shown that the proposed outage exponent is not only an accurate estimation of the outage probability for any number of subchannels, any SNR, and any target transmission rate, but also provides an easy way to compute the outage capacity, finite-SNR diversity-multiplexing tradeoff, and SNR gain. The asymptotic performance metrics, such as the delay-limited capacity, ergodic capacity, and diversity-multiplexing tradeoff can be directly obtained by letting the number of subchannels or SNR tends to infinity. Similar to Gallager's error exponent, a reliable function for parallel fading channels, which illustrates a fundamental relationship between the transmission reliability and efficiency, can also be defined from the outage exponent. Therefore, the proposed outage exponent provides a complete and comprehensive performance measure for parallel fading channels.

Introduction

Modern wireless communication systems, such as wide-band OFDM and MIMO, are typically modeled as parallel fading channels. Because these systems utilize a finite number of subchannels, they are inherently non-ergodic—the channel capacity fluctuates, making it impossible to guarantee a fixed rate over short intervals. Consequently, the primary performance metric for such systems is the outage probability: the probability that the instantaneous channel capacity falls below a target transmission rate.

Historically, calculating exact outage probability has been mathematically challenging. Existing approximations were generally limited to asymptotic regimes, such as an infinite number of subchannels or extremely high Signal-to-Noise Ratios (SNR). This paper addresses this gap by introducing the outage exponent, a unified framework that provides exponentially tight bounds for outage probability across all practical scenarios.

Methodology: The Outage Exponent Framework

The authors derive the outage exponent by splitting the analysis into two distinct regimes based on the target transmission rate ($R$) relative to the channel's ergodic capacity ($C$).

The multiplexing gain threshold $r_0(\gamma)$ delineates the regions where large deviations theory or Meijer's G-function are applied.
  1. Low-Rate Regime ($R < C$): The framework employs large deviations theory, specifically saddle-point approximations and Cramér's theorem, to bound the probability of rare capacity drops.
  2. High-Rate Regime ($R \ge C$): In this region, large deviations theory is insufficient. The authors instead utilize contour integration and the Meijer's G-function to compute the bounds.

A significant innovation of this approach is the isolation of the fading effect from the channel coding scheme. By assuming an ideal coding scheme, the outage exponent provides a simpler and tighter metric than traditional error exponents. The mathematical validity of these bounds was verified using synthetic Monte Carlo simulations.

Theoretical Contributions: The Reliable Function

The outage exponent serves as a "master framework" from which other critical asymptotic metrics—such as delay-limited capacity, ergodic capacity, and the diversity-multiplexing tradeoff (DMT)—can be directly derived.

A central theoretical insight is the definition of a reliable function for parallel fading channels, analogous to Gallager's error exponent in coding theory.

The reliable function $E(r; \gamma)$ illustrates the fundamental relationship between transmission reliability and multiplexing gain at various SNR levels.

This function demonstrates that the outage probability approaches zero exponentially as the number of subchannels increases, provided the target rate is strictly less than the ergodic capacity. It bridges the gap between finite-SNR realities and asymptotic theory, allowing for the calculation of the "SNR gain" required to achieve specific reliability targets.

Performance Evaluation and Results

The proposed outage exponent demonstrates near-perfect accuracy when compared to simulated outage probabilities across a wide range of SNRs (-20 dB to 40 dB) and subchannel counts ($L=2, 4, 6$).

Key findings include:

  • Superiority over Baselines: The framework significantly outperforms standard approximations, such as the Kaplan-Shamai upper bound. While older models often only captured the slope of the outage curve at high SNR, the outage exponent accurately predicts both the slope and the intercept at realistic SNRs.
  • DMT Convergence: The finite-SNR diversity-multiplexing tradeoff derived from the framework perfectly converges to theoretical asymptotic DMT values at high SNRs (>40 dB), validating the model's consistency across scales.

Conclusion and Future Work

While the outage exponent offers a powerful unified metric, its practical application faces computational challenges, particularly regarding the numerical evaluation of the Meijer's G-function in certain logarithmic cases. Future research is directed toward designing practical coding schemes—such as optimized $\mathbb{Z}^L$-lattices and permutation codes—that can achieve the theoretical performance limits established by this framework.

Summary updated

Abstract

The parallel fading channel, which consists of finite number of subchannels, is very important, because it can be used to formulate many practical communication systems. The outage probability, on the other hand, is widely used to analyze the relationship among the communication efficiency, reliability, SNR, and channel fading. To the best of our knowledge, the previous works only studied the asymptotic outage performance of the parallel fading channel which are only valid for a large number of subchannels or high SNRs. In this paper, a unified performance metric, which we shall refer to as the outage exponent, will be proposed. Our approach is mainly based on the large deviations theory and the Meijer's G-function. It is shown that the proposed outage exponent is not only an accurate estimation of the outage probability for any number of subchannels, any SNR, and any target transmission rate, but also provides an easy way to compute the outage capacity, finite-SNR diversity-multiplexing tradeoff, and SNR gain. The asymptotic performance metrics, such as the delay-limited capacity, ergodic capacity, and diversity-multiplexing tradeoff can be directly obtained by letting the number of subchannels or SNR tends to infinity. Similar to Gallager's error exponent, a reliable function for parallel fading channels, which illustrates a fundamental relationship between the transmission reliability and efficiency, can also be defined from the outage exponent. Therefore, the proposed outage exponent provides a complete and comprehensive performance measure for parallel fading channels.
BibTeX
@article{bai2013outage,
  author = {Bo Bai and Wei Chen and Khaled Ben Letaief and Zhigang Cao},
  title = {{Outage Exponent: A Unified Performance Metric for Parallel Fading Channels}},
  journal = {IEEE Trans. Information Theory},
  year = {2013},
  url = {https://openreview.net/forum?id=SNjPf86SCf},
}

Loading related content...

Figure