Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Reinforcement Learning-based Relay Selection for Cooperative WSNs in the Presence of Bursty Impulsive Noise

Hazem Barka, Md Sahabul Alam, Georges Kaddoum, Minh Au, Basile L. Agba

TL;DR

The paper tackles relay selection in cooperative WSNs with Decode-and-Forward relays under bursty impulsive noise with memory modeled as a Two-state Markov Gaussian process, focusing on Symbol-Error-Rate and battery fairness. It introduces two strategies: (i) a noise-state-aware Max-Min criterion augmented with residual battery metrics to improve SER and balance energy consumption, and (ii) a reinforcement-learning approach using policy-gradient (REINFORCE) to learn relay selection in an MDP where states encode channel gains, impulse activity, and battery levels. Experimental results show the modified Max-Min can approach AWGN-like SER while achieving near-equal remaining energy among relays, and the RL method can match AWGN/TSMG benchmarks and exploit impulsive-noise patterns for improved performance and fairness. The methods are applicable to multi-relay scenarios and can accommodate additional WSN constraints, enhancing reliability and energy efficiency in practical smart-grid-like environments.

Abstract

The problem of relay selection is pivotal in the realm of cooperative communication. However, this issue has not been thoroughly examined, particularly when the background noise is assumed to possess an impulsive characteristic with consistent memory as observed in smart grid communications and some other wireless communication scenarios. In this paper, we investigate the impact of this specific type of noise on the performance of cooperative Wireless Sensor Networks (WSNs) with the Decode and Forward (DF) relaying scheme, considering Symbol-Error-Rate (SER) and battery power consumption fairness across all nodes as the performance metrics. We introduce two innovative relay selection methods that depend on noise state detection and the residual battery power of each relay. The first method encompasses the adaptation of the Max-Min criterion to this specific context, whereas the second employs Reinforcement Learning (RL) to surmount this challenge. Our empirical outcomes demonstrate that the impacts of bursty impulsive noise on the SER performance can be effectively mitigated and that a balance in battery power consumption among all nodes can be established using the proposed methods.

Reinforcement Learning-based Relay Selection for Cooperative WSNs in the Presence of Bursty Impulsive Noise

TL;DR

The paper tackles relay selection in cooperative WSNs with Decode-and-Forward relays under bursty impulsive noise with memory modeled as a Two-state Markov Gaussian process, focusing on Symbol-Error-Rate and battery fairness. It introduces two strategies: (i) a noise-state-aware Max-Min criterion augmented with residual battery metrics to improve SER and balance energy consumption, and (ii) a reinforcement-learning approach using policy-gradient (REINFORCE) to learn relay selection in an MDP where states encode channel gains, impulse activity, and battery levels. Experimental results show the modified Max-Min can approach AWGN-like SER while achieving near-equal remaining energy among relays, and the RL method can match AWGN/TSMG benchmarks and exploit impulsive-noise patterns for improved performance and fairness. The methods are applicable to multi-relay scenarios and can accommodate additional WSN constraints, enhancing reliability and energy efficiency in practical smart-grid-like environments.

Abstract

The problem of relay selection is pivotal in the realm of cooperative communication. However, this issue has not been thoroughly examined, particularly when the background noise is assumed to possess an impulsive characteristic with consistent memory as observed in smart grid communications and some other wireless communication scenarios. In this paper, we investigate the impact of this specific type of noise on the performance of cooperative Wireless Sensor Networks (WSNs) with the Decode and Forward (DF) relaying scheme, considering Symbol-Error-Rate (SER) and battery power consumption fairness across all nodes as the performance metrics. We introduce two innovative relay selection methods that depend on noise state detection and the residual battery power of each relay. The first method encompasses the adaptation of the Max-Min criterion to this specific context, whereas the second employs Reinforcement Learning (RL) to surmount this challenge. Our empirical outcomes demonstrate that the impacts of bursty impulsive noise on the SER performance can be effectively mitigated and that a balance in battery power consumption among all nodes can be established using the proposed methods.
Paper Structure (14 sections, 4 equations, 10 figures, 1 table, 2 algorithms)

This paper contains 14 sections, 4 equations, 10 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. System Model
  3. Signal Model
  4. Assumptions on the Channel Coefficients
  5. Assumptions on the Noise
  6. The Max-Min relay Selection Criterion
  7. Impact of Bursty Impulsive Noise on the Conventional Max-Min Criterion
  8. Proposed Max-Min Relay Selection Criterion Adapted to Bursty Impulsive Noise Environments with Battery Fairness
  9. Method
  10. Numerical Results and Interpretations
  11. Reinforcement Learning Based Relay Selection in the Presence of Bursty Impulsive Noise
  12. The Proposed Method
  13. Numerical Results and Interpretations
  14. Conclusion

Figures (10)

  • Figure 1: Illustration of the sensors' repartition.
  • Figure 2: SER performance using the conventional max-min criterion for a QPSK modulated symbols under TSMG and AWGN noise and slow Rayleigh fading.
  • Figure 3: SER performance using the conventional Max-Min criterion for a QPSK modulated symbol under TSMG and AWGN noise and fast Rayleigh fading ($T_c = T_s$).
  • Figure 4: The received signal power by symbol $y_{S R_{m}, k}^2$ for a frame of size 1000 symbols under the considered bursty impulsive noise scenario.
  • Figure 5: Remaining battery level for each relay node after the simulated communication.
  • ...and 5 more figures