Due to the advantages of high mobility and line-of-sight transmission,unmanned aerial vehi-cles(UAVs)equipped with mobile edge computing(MEC)servers can effectively reduce the computa-tional burden and task delay of g...Due to the advantages of high mobility and line-of-sight transmission,unmanned aerial vehi-cles(UAVs)equipped with mobile edge computing(MEC)servers can effectively reduce the computa-tional burden and task delay of ground users(GUs).However,the offloading data from GU to UAV is vul-nerable to be eavesdropped by malicious users in the network.Thus,this paper proposes a secure coopera-tive offloading scheme in a multi-UAV-assisted MEC network,where each UAV has the capability to par-tially distributing the tasks to other idle UAVs.Specif-ically,we first model the task offloading decision pro-cess of GUs based on the multi-agent Markov Deci-sion Process(MDP)framework.Then we optimize the offloading decision of GUs by adopting multi-agent deep determined policy gradient(MADDPG)to min-imize the overall system latency for task processing and computation offloading.Simulation results verify that the proposed cooperative offloading scheme can effectively reduce the system latency compared with the benchmark.展开更多
The utilization of traffic information received from intelligent vehicle highway systems(IVHS) to plan velocity and split output power for multi-source vehicles is currently a research hotspot. However, it is an open ...The utilization of traffic information received from intelligent vehicle highway systems(IVHS) to plan velocity and split output power for multi-source vehicles is currently a research hotspot. However, it is an open issue to plan vehicle velocity and distribute output power between different supply units simultaneously due to the strongly coupling characteristic of the velocity planning and the power distribution. To address this issue, a flexible predictive power-split control strategy based on IVHS is proposed for electric vehicles(EVs) equipped with battery-supercapacitor system(BSS). Unlike hierarchical strategies to plan vehicle velocity and distribute output power separately, a monolayer model predictive control(MPC) method is employed to optimize them online at the same time. Firstly, a flexible velocity planning strategy is designed based on the signal phase and time(SPAT) information received from IVHS and then the Pontryagin’s minimum principle(PMP) is adopted to formulate the optimal control problem of the BSS. Then, the flexible velocity planning strategy and the optimal control problem of BSS are embedded into an MPC framework, which is online solved using the shooting method in a fashion of receding horizon. Simulation results verify that the proposed strategy achieves a superior performance compared with the hierarchical strategy in terms of transportation efficiency, battery capacity loss, energy consumption and computation time.展开更多
基金supported in part by the Zhejiang Provincial Natural Science Foundation of China under Grant LR25F010003 and ZCLZ25F0102in part by the National Natural Science Foundation of China under Grant 62271447,62171412,61871348 and 62471090+3 种基金in part by the Fundamental Research Funds for the Provincial Universities of Zhejiang under Grant RF-C2023008in part by the Central University Fund under Grant ZYGX2024Z016in part by the Sichuan Provincial Natural Science Foundation Project under Grant 23NSFSC0422in part by the Intelligent Terminal Key Laboratory of Sichuan Province under Grant SCITLAB-20005.
文摘Due to the advantages of high mobility and line-of-sight transmission,unmanned aerial vehi-cles(UAVs)equipped with mobile edge computing(MEC)servers can effectively reduce the computa-tional burden and task delay of ground users(GUs).However,the offloading data from GU to UAV is vul-nerable to be eavesdropped by malicious users in the network.Thus,this paper proposes a secure coopera-tive offloading scheme in a multi-UAV-assisted MEC network,where each UAV has the capability to par-tially distributing the tasks to other idle UAVs.Specif-ically,we first model the task offloading decision pro-cess of GUs based on the multi-agent Markov Deci-sion Process(MDP)framework.Then we optimize the offloading decision of GUs by adopting multi-agent deep determined policy gradient(MADDPG)to min-imize the overall system latency for task processing and computation offloading.Simulation results verify that the proposed cooperative offloading scheme can effectively reduce the system latency compared with the benchmark.
基金supported by the National Natural Science Foundation of China (62173303)the Fundamental Research for the Zhejiang P rovincial Universities (RF-C2020003)。
文摘The utilization of traffic information received from intelligent vehicle highway systems(IVHS) to plan velocity and split output power for multi-source vehicles is currently a research hotspot. However, it is an open issue to plan vehicle velocity and distribute output power between different supply units simultaneously due to the strongly coupling characteristic of the velocity planning and the power distribution. To address this issue, a flexible predictive power-split control strategy based on IVHS is proposed for electric vehicles(EVs) equipped with battery-supercapacitor system(BSS). Unlike hierarchical strategies to plan vehicle velocity and distribute output power separately, a monolayer model predictive control(MPC) method is employed to optimize them online at the same time. Firstly, a flexible velocity planning strategy is designed based on the signal phase and time(SPAT) information received from IVHS and then the Pontryagin’s minimum principle(PMP) is adopted to formulate the optimal control problem of the BSS. Then, the flexible velocity planning strategy and the optimal control problem of BSS are embedded into an MPC framework, which is online solved using the shooting method in a fashion of receding horizon. Simulation results verify that the proposed strategy achieves a superior performance compared with the hierarchical strategy in terms of transportation efficiency, battery capacity loss, energy consumption and computation time.
