Design a MAC Protocol to Improve Data Storage and Transmission using Reinforcement Learning

[ 31 May 2023 | vol. 16 | no. 1 | pp. 11-22 ]

About Authors:

Yu-Chen Hu
-Providence University, Taiwan

Abstract:

In this paper, we present a novel algorithm for data storage and transmission that leverages Reinforcement Learning (RL) techniques to optimize efficiency within wireless sensor networks (WSNs). Our proposed protocol effectively addresses the challenges associated with data management by implementing a priority-based framework for both data storage and transmission. The core of our algorithm hinges on the categorization of incoming data into two distinct types: Event Data and Normal Data. Event Data represents critical or time-sensitive information, while Normal Data encompasses less urgent content. By employing the RL-Classifier, our method assigns priority levels to these data types, thereby organizing them within a transmission queue based on their urgency and importance. A fundamental aspect of our approach is the implementation of priority-aware transmission. During the data transmission phase, we incorporate differentiated transmission probabilities that align with the assigned priority levels of the data queued for sending. This ensures that higher-priority data is transmitted more reliably and with greater urgency, thereby reducing latency and improving overall network performance. To evaluate the efficacy of our proposed protocol, we conducted a comparative analysis against the existing QAML-MAC framework, which was designed to enhance the efficiency of data storage and transmission within WSNs. Our experiments demonstrate a significant improvement in performance metrics when utilizing our RL-based algorithm, emphasizing its potential to enhance data handling capabilities in wireless sensor networks. This paper ultimately underscores the benefits of integrating reinforcement learning into wireless communication systems for optimized data prioritization and management.

Keywords:

Wireless Sensor Networks, MAC protocol, Priority aware, Reinforce-learning, Throughput, Quality of Service

 

About this Article: