He is a postdoctoral fellow at the College of Mechatronics and Control Engineering, Shenzhen University, Shenzhen 518060, China.
I am delighted to announce that our paper entitled "Multi-Agent DRL-Based Energy Harvesting for Freshness of Data in UAV-Assisted Wireless Sensor Networks" has been published in the IEEE Transactions on Network and Service Management (Early Access) (Q1,IF=4.7). First of all, I would like to thank one of our co-authors and my dear friends, Dr. Mohammed Seid, Ph.D and Nahom Abishu Hayla, for their countless contributions to our paper. Congratulations to all authors, Prof. Supeng Leng, Dr. Maged Refat , Prof. Aiman Erbad, and Prof. Mohsen Guizani. We have introduced a novel multi-access edge computing (MEC)-integrated UAV-supported wireless sensor network (WSN) with a laser technology-based energy harvesting (EH) system that makes the UAV act as a flying energy charger to address these issues. This work aims to minimize the age of information (AoI) and improve energy efficiency by jointly optimizing the UAV trajectories, EH, task scheduling, and data offloading. The joint optimization problem is formulated as a Markov decision process (MDP) and then transformed into a stochastic game model to handle the complexity and dynamics of the environment. We employ a multi-agent deep Q-network (MADQN) algorithm to solve the formulated optimization problem. With the MADQN algorithm, UAVs can determine the best data collection and EH decisions to minimize their energy consumption and efficiently collect data from multiple SNs, leading to reduced AoI and improved energy efficiency.
Founder of SmythOS.com | AI Multi-Agent Orchestration ▶️
6moFascinating UAV-enabled WSN optimization. MDP formulation and MADQN algorithm handling complexities sounds ingenious. Kudos on the impactful Q1 IEEE publication.