AoI-Optimal Data Collection, Offloading, and Migration in Mobile Edge Networks

Abstract

With the explosive development of Internet of Things (IoT) devices, edge sensors are deployed densely to monitor the environment. The status data can be sampled by the edge sensors and be transmitted to the mobile end device, such as an unmanned aerial vehicle (UAV), for further execution or offloading. To make an alert decision on time, the UAV requires to frequently collect and rapidly process the status data. The data freshness can be measured by the age of information (AoI). The UAV needs to design a flying trajectory to collect data from a large area. In addition, to satisfy the real-time requirement of the computation-intensive tasks, the UAV may offload the tasks to nearby edge servers and migrate service due to its mobility. In this paper, there are several challenges in the AoI-optimal UAV-assisted edge scenario: the freshness challenge, the collection challenge, the offloading, and the migration challenge. To minimize the time between two adjacent sampling for all sensors, we propose the AoI-optimal UAV Collection Control solution (AUCC). Specifically, to minimize round collection time, we propose the Dynamic programming-based Task Collection algorithm (DTC); to minimize round execution time, we propose the Lyapunov optimization-based Accuracy queue Control algorithm (LAC). Simulation results demonstrate that our AUCC solution achieves superior performance from the perspective of AoI, overtime task number, and accuracy.