Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DeepNP: Deep Learning-Based Noise Prediction for Ultra-Reliable Low-Latency Communications

Adina Waxman, Nir Shlezinger, Alejandro Cohen

Abstract

Adaptive network coding schemes provide a promising approach to bridging the gap between high data rates and low delay in real-time streaming applications. However, their effectiveness often relies on accurate channel prediction, which is typically based on delayed feedback and is especially challenging when the underlying channel model is unknown. To address this, we introduce a novel integration of network coding with a channel-agnostic, Deep learning-based Noise Prediction algorithm (DeepNP). Unlike traditional estimators, DeepNP predicts statistical noise rates rather than instantaneous noise realizations, significantly simplifying the prediction task while enhancing coding performance. DeepNP is designed to operate with both binary (e.g., acknowledgments) and continuous-valued (e.g., Signal-to-Noise Ratio, SNR) feedback. We incorporate DeepNP into the Adaptive and Causal Random Linear Network Coding (AC-RLNC) framework to jointly optimize throughput and in-order delivery delay. Two variants are proposed: (i) Erasure-Rate DeepNP (ER-DeepNP), which serves as a transport-layer noise predictor and achieves in a numerical study up to a 2x reduction in mean and maximum delay with less than 0.1 loss in throughput compared to statistic-based estimators, under Round-Trip Time (RTT) up to 40 time slots and erasure rates up to 60%; and (ii) Cross-Layer DeepNP (CL-DeepNP), which dynamically adjusts the SNR threshold to maintain high physical layer code rates while achieving low transport-layer erasure rates. This yields, in the presented numerical study, a 25% throughput gain over fixed-threshold approaches. Our results demonstrate that DeepNP enables robust, model-free noise prediction, making adaptive network coding more viable in practical, feedback-limited communication scenarios.

DeepNP: Deep Learning-Based Noise Prediction for Ultra-Reliable Low-Latency Communications

Abstract

Adaptive network coding schemes provide a promising approach to bridging the gap between high data rates and low delay in real-time streaming applications. However, their effectiveness often relies on accurate channel prediction, which is typically based on delayed feedback and is especially challenging when the underlying channel model is unknown. To address this, we introduce a novel integration of network coding with a channel-agnostic, Deep learning-based Noise Prediction algorithm (DeepNP). Unlike traditional estimators, DeepNP predicts statistical noise rates rather than instantaneous noise realizations, significantly simplifying the prediction task while enhancing coding performance. DeepNP is designed to operate with both binary (e.g., acknowledgments) and continuous-valued (e.g., Signal-to-Noise Ratio, SNR) feedback. We incorporate DeepNP into the Adaptive and Causal Random Linear Network Coding (AC-RLNC) framework to jointly optimize throughput and in-order delivery delay. Two variants are proposed: (i) Erasure-Rate DeepNP (ER-DeepNP), which serves as a transport-layer noise predictor and achieves in a numerical study up to a 2x reduction in mean and maximum delay with less than 0.1 loss in throughput compared to statistic-based estimators, under Round-Trip Time (RTT) up to 40 time slots and erasure rates up to 60%; and (ii) Cross-Layer DeepNP (CL-DeepNP), which dynamically adjusts the SNR threshold to maintain high physical layer code rates while achieving low transport-layer erasure rates. This yields, in the presented numerical study, a 25% throughput gain over fixed-threshold approaches. Our results demonstrate that DeepNP enables robust, model-free noise prediction, making adaptive network coding more viable in practical, feedback-limited communication scenarios.

Paper Structure

This paper contains 24 sections, 27 equations, 21 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Preliminaries
  3. Communication Channel Model
  4. Problem Formulation
  5. Metrics
  6. Objective
  7. Challenges
  8. Preliminaries
  9. Deep Learning Aided Noise Prediction
  10. DeepNP - Architecture
  11. DeepNP - Training
  12. Data
  13. Loss
  14. ER-DeepNP
  15. CL-DeepNP
  16. ...and 9 more sections

Figures (21)

  • Figure 1: System Model. The sender uses the feedback to estimate the channel rate and encodes packets using acrlnc. It may adjust the phy code rate for the next transmission.
  • Figure 2: DeepNP architectures for binary and snr feedback types. (a) For the binary feedback type, each block's input includes the output erasure probability. (b) For the snr feedback type, the feedback is concatenated with the predicted snr (i.e., the output of the fully-connected layer).
  • Figure 3: $U$ parallel DeepNP models extended to $\rm RTT$$+n$ building block. Each model has its own constant threshold.
  • Figure 4: DeepNP extended to $\rm RTT$$+n$ building blocks. The first $\rm RTT$ blocks' threshold is taken from the input, While the last $n$ blocks' threshold is generated from a separate neural block.
  • Figure 5: Simulation snapshot: visualizing SINR with a 5dB SINR threshold in the top row (red line), accompanied by the corresponding (binary) erasure series in the bottom row.
  • ...and 16 more figures