Learning medium access control (MAC) protocols through the multi-agent reinforcement learning (MARL) paradigm has been applied in simple wireless networks, enabling efficient goal-oriented channel access policies. However, no studies in the literature have considered industrial environments, which could benefit significantly from MAC protocols tailored to their stringent requirements, such as energy efficiency for battery-constrained devices. In this paper, we define a new MARL framework nestled in an industrial environment to provide energy-efficient MAC protocols. Moreover, the proposed MARL framework is feasible, i.e., it enables practical implementations of the training procedures in real-world scenarios. This feature allows learned MAC protocols to be tailor-made for the unique characteristics of the industrial scenario, in terms of traffic pattern and channel conditions. We evaluate the performance of the proposed MARL framework by comparing it to both a benchmark MARL-based solution and a conventional grant-based MAC protocol, neither of which is specifically tailored for industrial environments. Simulation results from the testing of the learned MAC protocols demonstrate that literature approaches are partially successful even in simple industrial scenarios, while the proposed framework effectively achieves its objectives in various industrial scenarios.

Learning Energy-Efficient MAC Protocols Using MARL in Industrial Networks

Miuccio L.;Riolo S.;Panno D.
2025-01-01

Abstract

Learning medium access control (MAC) protocols through the multi-agent reinforcement learning (MARL) paradigm has been applied in simple wireless networks, enabling efficient goal-oriented channel access policies. However, no studies in the literature have considered industrial environments, which could benefit significantly from MAC protocols tailored to their stringent requirements, such as energy efficiency for battery-constrained devices. In this paper, we define a new MARL framework nestled in an industrial environment to provide energy-efficient MAC protocols. Moreover, the proposed MARL framework is feasible, i.e., it enables practical implementations of the training procedures in real-world scenarios. This feature allows learned MAC protocols to be tailor-made for the unique characteristics of the industrial scenario, in terms of traffic pattern and channel conditions. We evaluate the performance of the proposed MARL framework by comparing it to both a benchmark MARL-based solution and a conventional grant-based MAC protocol, neither of which is specifically tailored for industrial environments. Simulation results from the testing of the learned MAC protocols demonstrate that literature approaches are partially successful even in simple industrial scenarios, while the proposed framework effectively achieves its objectives in various industrial scenarios.
2025
6G
feasibility
industrial networks
MAC
multi-agent reinforcement learning
protocol learning
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11769/686631
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact