Researchers from the Beijing University of Posts and Telecommunications, including Siqi Mu, Shuo Wen, Yang Lu, Ruihong Jiang, and Bo Ai, have developed a novel approach to optimize the use of unmanned aerial vehicles (UAVs) in the Internet of Medical Things (IoMT). Their work aims to improve real-time biomedical edge computing services for wireless body area network (WBAN) users, which are networks of wearable and implantable medical devices.
The researchers have tackled the challenge of dynamic task offloading and UAV flight trajectory optimization to minimize the weighted average task completion time for all WBAN users, while also considering the energy consumption constraints of the UAVs. To achieve this, they have established an embodied AI-enhanced IoMT edge computing framework. This framework utilizes a hierarchical multi-scale Transformer-based user trajectory prediction model, which is based on the users’ historical trajectory traces captured by the UAVs.
The researchers have also designed a prediction-enhanced deep reinforcement learning (DRL) algorithm that integrates predicted users’ mobility information. This algorithm is used to intelligently optimize UAV flight trajectories and task offloading decisions. The effectiveness of the proposed methods was demonstrated using real-world movement traces and simulation results, which showed that the methods outperformed existing benchmarks.
The practical applications of this research for the energy sector include improving the efficiency of UAVs used for monitoring and maintaining energy infrastructure, such as power lines and wind turbines. By optimizing flight trajectories and task offloading, UAVs can complete inspections and maintenance tasks more quickly and with less energy consumption. This can lead to cost savings and reduced downtime for energy infrastructure, ultimately contributing to a more reliable and sustainable energy system. The research was published in the IEEE Internet of Things Journal.
This article is based on research available at arXiv.