Age of Local Information for Fusion Freshness in Internet of Things

In recent years, the applications of Internet of Things (IoT) have been extensively explored. Among these applications, information fusion based on probability theories or Artificial Intelligence (AI) plays a foundational role. However, to the best of our knowledge, few works consider the scheduling of the IoT nodes that collect redundant data to promise the freshness of the fused information. Age of Information (AoI) is a metric that characterizes the obsolescence of information. Although AoI works well in single-source updating systems, it is hardly applicable to a multi-source system setting such as IoT. In this paper, we propose a new metric called Age of Local Information (AoLI) to illustrate the obsolescence of data collected by each node of IoT used in fusion. We first simulate AoLI in an FCFS queue and show that the optimal arrival rate that minimizes the mean AoI does not necessarily result in a quicker information fusion. Instead, the arrival rate that minimizes the maximal AoI performs better. We then propose a scheduling policy called the Fusion Greedy policy to schedule the IoT nodes in a discrete manner under an FCFS queue of length 1. Through simulation, we demonstrate that the proposed policy outperforms traditional policies such as Whittle’s index policy or the AoI Greedy policy.