Research On Edge Collaborative Processing Mechanism Of Internet Of Things Data For Multi-UAV Scenarios

With the rapid development of the Internet of Things(Io T),terminal devices are facing increasingly computational-intensive tasks,which cannot be processed effectively due to limitations in processing capacity and battery life.Mobile Edge Computing(MEC)and Unmanned Aerial Vehicle(UAV)have been widely used to provide enhanced coverage and computing services to ground-based Io T systems.Terminal devices can offload tasks to the MEC server on UAV to obtain low-latency services,however,it is challenging for a single UAV to process the massive computing tasks transmitted by terminal devices.Based on the shortcomings of traditional mobile edge computing,we propose a threelayer Io T data processing architecture with multi-UAV collaboration and design a multiUAV collaborative offloading strategy.The proposed architecture can provide adequate computing offloading services for terminal devices in two scenarios,with cellular network coverage(centralized)and without cellular network coverage(distributed).The main contributions of this article are summarized as follows:1.For the centralized scenario,we first construct the optimization problem of minimizing the delay in task completion,and use the Lyapunov-optimized online collaborative algorithm to decouple the energy constraints of UAVs,then we apply the proposed centralized multi-UAV cooperative offloading algorithm to obtain optimal cooperative offloading strategy between UAV and cloud,finally,the performance of Lyapunov optimization algorithm is verified by rigorous mathematical analysis;2.For the distributed scenario,we built a non-cooperative game model of multi-UAV,and use the best response algorithm to obtaining the offloading strategy of the UAV,then we solve the Nash equilibrium based on our distributed multi-UAV cooperative offloading algorithm,thus the cooperative offloading strategy profile between UAVs is obtained;3.Multi-dimensional simulation and comparison of the proposed algorithms were carried out in both two scenarios,then the impact of Lyapunov control parameters and task arrival rate of terminal devices on the performance of the algorithm is studied.Simulation results show that the proposed algorithm can achieve a lower task completion delay compared with other benchmark algorithms.