Optimal Mobility and Communication Strategy to Maximize the Value of Information in IoT Networks

TL;DR

This paper develops a reinforcement learning approach to optimize mobility and communication strategies in industrial IoT networks, aiming to maximize the value of information for real-time monitoring and control.

Contribution

It formulates a joint path planning and user scheduling problem as an MDP and proposes an efficient RL-based solution to enhance information value in resource-constrained IoT systems.

Findings

The proposed method effectively captures data usefulness over time and space.

It leverages data correlation to improve performance in resource-limited networks.

Numerical results demonstrate the approach's ability to maximize network-level VoI.

Abstract

The Internet of Things (IoT) is an emerging next-generation technology in the fourth industrial revolution. In industrial IoT networks, sensing devices are largely deployed to monitor various types of physical processes. They are required to transmit the collected data in a timely manner to support real-time monitoring, control and automation. The timeliness of information is very important in such systems. Recently, an information-theoretic metric named the "value of information" (VoI) has been proposed to measure the usefulness of information. In this work, we consider an industrial IoT network with a set of heterogeneous sensing devices and an intelligent mobile entity. The concept of the value of information is applied to study a joint path planning and user scheduling optimisation problem. We aim to maximise the network-level VoI under mobility and communication constraints. We formulate this problem as a Markov decision process (MDP), and an efficient algorithm based on reinforcement learning is proposed to solve this problem. Through numerical results, we show that the proposed method is able to capture the usefulness of data from both time and space dimensions. By exploiting the correlation property of the data source, the proposed method is suitable for applications in resource-limited networks.