Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Ultra-High Reliability by Predictive Interference Management Using Extreme Value Theory

Fateme Salehi, Aamir Mahmood, Sinem Coleri, Mikael Gidlund

TL;DR

A risk-sensitive approach based on extreme value theory (EVT) to predict the signal-to-interference-plus-noise ratio (SINR) for efficient resource allocation in URLLC systems is proposed, and is sample-efficient, able to predict interference effectively with minimal training data.

Abstract

Ultra-reliable low-latency communications (URLLC) require innovative approaches to modeling channel and interference dynamics, extending beyond traditional average estimates to encompass entire statistical distributions, including rare and extreme events that challenge achieving ultra-reliability performance regions. In this paper, we propose a risk-sensitive approach based on extreme value theory (EVT) to predict the signal-to-interference-plus-noise ratio (SINR) for efficient resource allocation in URLLC systems. We employ EVT to estimate the statistics of rare and extreme interference values, and kernel density estimation (KDE) to model the distribution of non-extreme events. Using a mixture model, we develop an interference prediction algorithm based on quantile prediction, introducing a confidence level parameter to balance reliability and resource usage. While accounting for the risk sensitivity of interference estimates, the prediction outcome is then used for appropriate resource allocation of a URLLC transmission under link outage constraints. Simulation results demonstrate that the proposed method outperforms the state-of-the-art first-order discrete-time Markov chain (DTMC) approach by reducing outage rates up to 100-fold, achieving target outage probabilities as low as \(10^{-7}\). Simultaneously, it minimizes radio resource usage \(\simnot15 \%\) compared to DTMC, while remaining only \(\simnot20 \%\) above the optimal case with perfect interference knowledge, resulting in significantly higher prediction accuracy. Additionally, the method is sample-efficient, able to predict interference effectively with minimal training data.

Ultra-High Reliability by Predictive Interference Management Using Extreme Value Theory

TL;DR

A risk-sensitive approach based on extreme value theory (EVT) to predict the signal-to-interference-plus-noise ratio (SINR) for efficient resource allocation in URLLC systems is proposed, and is sample-efficient, able to predict interference effectively with minimal training data.

Abstract

Ultra-reliable low-latency communications (URLLC) require innovative approaches to modeling channel and interference dynamics, extending beyond traditional average estimates to encompass entire statistical distributions, including rare and extreme events that challenge achieving ultra-reliability performance regions. In this paper, we propose a risk-sensitive approach based on extreme value theory (EVT) to predict the signal-to-interference-plus-noise ratio (SINR) for efficient resource allocation in URLLC systems. We employ EVT to estimate the statistics of rare and extreme interference values, and kernel density estimation (KDE) to model the distribution of non-extreme events. Using a mixture model, we develop an interference prediction algorithm based on quantile prediction, introducing a confidence level parameter to balance reliability and resource usage. While accounting for the risk sensitivity of interference estimates, the prediction outcome is then used for appropriate resource allocation of a URLLC transmission under link outage constraints. Simulation results demonstrate that the proposed method outperforms the state-of-the-art first-order discrete-time Markov chain (DTMC) approach by reducing outage rates up to 100-fold, achieving target outage probabilities as low as . Simultaneously, it minimizes radio resource usage compared to DTMC, while remaining only above the optimal case with perfect interference knowledge, resulting in significantly higher prediction accuracy. Additionally, the method is sample-efficient, able to predict interference effectively with minimal training data.
Paper Structure (13 sections, 13 equations, 5 figures, 1 table)

This paper contains 13 sections, 13 equations, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. System Model
  3. The Proposed Interference Prediction and Resource Allocation Method
  4. Interference Prediction
  5. GPD for the Tail Distribution
  6. KDE for the Bulk Distribution
  7. Resource Allocation
  8. Implementation Perspective
  9. Numerical Results
  10. Performance Under Different Confidence Levels
  11. Performance Under the Resource Limit
  12. Performance Under Different Number of Training Samples
  13. Conclusion

Figures (5)

  • Figure 1: Illustration of the mixture-based interference distribution modeling approach combining kernel density estimation (KDE) and generalized Pareto distribution (GPD). The blue dots and red dots represent data points used for density estimation, respectively in the bulk region and in the tail region.
  • Figure 2: Comparison of achieved outage vs. target outage probability for the proposed mixture-based method against DTMC-based method for different confidence levels $\eta = \{0.95,0.99\}$.
  • Figure 3: Comparison of resource usage ratio vs. target outage probability for the proposed mixture-based method against DTMC-based method for different confidence levels $\eta = \{0.95,0.99\}$.
  • Figure 4: Achieved outage and resource usage ratio vs. target outage probability under blocklength limit $M=10^5$ channel uses and confidence level $\eta=0.99$. The dashed lines indicate the $99.99$th percentile of the achieved outage.
  • Figure 5: Comparison of the achieved outage vs. the number of training samples for the proposed mixture-based method against DTMC-based method for $\eta=0.99$ and different target outages $\epsilon = \{10^{-5},10^{-7}\}$.