Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ML Framework for Wireless MAC Protocol Design

Navid Keshtiarast, Marina Petrova

TL;DR

The paper addresses the need for adaptive MAC protocols in next-generation wireless networks to satisfy diverse QoS requirements. It introduces a DRL-based framework that decomposes IEEE 802.11ac into atomic MAC blocks and uses PPO to learn application-specific channel access policies. Through detailed OMNeT++/INET simulations, the learned MAC outperforms legacy 802.11ac in throughput and latency across varied network sizes and traffic patterns. The work provides an open-source framework and demonstrates the feasibility of automated, environment-aware MAC synthesis for future WLANs.

Abstract

Adaptivity, reconfigurability and intelligence are key features of the next-generation wireless networks to meet the increasingly diverse quality of service (QoS) requirements of the future applications. Conventional protocol designs, however, struggle to provide flexibility and agility to changing radio environments, traffic types and different user service requirements. In this paper, we explore the potential of deep reinforcement learning (DRL), in particular Proximal Policy Optimization (PPO), to design and configure intelligent and application-specific medium access control (MAC) protocols. We propose a framework that enables the addition, removal, or modification of protocol features to meet individual application needs. The DRL channel access policy design empowers the protocol to adapt and optimize in accordance with the network and radio environment. Through extensive simulations, we demonstrate the superior performance of the learned protocols over legacy IEEE 802.11ac in terms of throughput and latency.

ML Framework for Wireless MAC Protocol Design

TL;DR

The paper addresses the need for adaptive MAC protocols in next-generation wireless networks to satisfy diverse QoS requirements. It introduces a DRL-based framework that decomposes IEEE 802.11ac into atomic MAC blocks and uses PPO to learn application-specific channel access policies. Through detailed OMNeT++/INET simulations, the learned MAC outperforms legacy 802.11ac in throughput and latency across varied network sizes and traffic patterns. The work provides an open-source framework and demonstrates the feasibility of automated, environment-aware MAC synthesis for future WLANs.

Abstract

Adaptivity, reconfigurability and intelligence are key features of the next-generation wireless networks to meet the increasingly diverse quality of service (QoS) requirements of the future applications. Conventional protocol designs, however, struggle to provide flexibility and agility to changing radio environments, traffic types and different user service requirements. In this paper, we explore the potential of deep reinforcement learning (DRL), in particular Proximal Policy Optimization (PPO), to design and configure intelligent and application-specific medium access control (MAC) protocols. We propose a framework that enables the addition, removal, or modification of protocol features to meet individual application needs. The DRL channel access policy design empowers the protocol to adapt and optimize in accordance with the network and radio environment. Through extensive simulations, we demonstrate the superior performance of the learned protocols over legacy IEEE 802.11ac in terms of throughput and latency.
Paper Structure (11 sections, 6 equations, 5 figures, 2 tables, 1 algorithm)

This paper contains 11 sections, 6 equations, 5 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Motivation
  3. Related work
  4. Contribution
  5. System Setup
  6. Deep reinforcement Learning
  7. Objective: Maximize Mean throughput
  8. Markov decision process (MDP)
  9. Proximal policy optimization (PPO) for designing MAC protocol
  10. Performance Evaluation
  11. Conclusions

Figures (5)

  • Figure 1: System architecture
  • Figure 2: IEEE 802.11 DCF protocol with all MAC functions
  • Figure 3: Simulation and Evaluation Setup
  • Figure 4: Learning rate for A2C and PPO
  • Figure 5: Comparison of mean throughput and delay results with IEEE 802.11ac