| In the future B5G/6G intelligent mobile networks,a large number of computation-intensive and high-energy applications make edge computing,where mobile terminals migrate tasks to edge servers for processing,widely used.However,mobile devices are usually limited in terms of battery capacity and computing power due to physical size constraints.A large number of computing tasks during edge computing task migration will accelerate the energy consumption of terminals and shorten the battery life of mobile devices.Therefore,how to migrate tasks efficiently and reduce the energy consumption of mobile terminals has been an open issue in edge computing research.The existing research mostly focused on the task migration optimization problem in a single scenario of sub-6GHz or millimeter wave.While the future network will be a scenario of mixed application of high-rate millimeter wave and long-range sub-6GHz,which is one of the very important candidate technologies for 6G.Therefore,this paper analyzes the characteristics of sub-6GHz and millimeter wave for task migration in terms of energy efficiency,transmission range and hardware complexity.And this paper proposes an uplink sub-6GHz and downlink millimeter wave task migration method,which based on joint uplink and downlink user clustering.The specific work is as follows:1.To solve the problem of low energy efficiency in traditional edge computing systems,we propose an uplink sub-6GHz and downlink millimeter wave algorithm for task migration.During task migration,uplink task transmission is limited by the low transmit power,small battery capacity and low computing power of mobile terminals.Therefore,we design the uplink migration scheme based on sub-6GHz multi-user orthogonal frequency division multiplexing(OFDM)to ensure energy-efficient and long-distance task migration at low signal-to-noise ratio(SNR)in mobile terminals.On the other hand,the computational results back-propagation is supported by the high downlink transmit power,sufficient energy and high hardware complexity of the base station.Therefore,we design a scheme based on millimeter wave massive multiple-input multiple-output(Massive MIMO)non-orthogonal multiple access(NOMA)for downlink high SNR conditions with energy harvesting.We use simultaneous wireless information and power transfer(SWIPT)technology to harvest energy to the mobile terminals while achieving high data rate.Simulation results show that the proposed uplink sub-6GHz and downlink millimeter wave task migration algorithm enables the mobile terminals to achieve higher energy efficiency and lower latency than conventional task migration schemes.2.To solve the problem of high complexity and low energy efficiency caused by the separation of uplink sub-6GHz MIMO user pairing and downlink millimeter wave user clustering in uplink sub-6GHz and downlink millimeter wave edge computing systems,this paper proposes a task migration algorithm based on joint uplink and downlink user clustering.In order to reduce computational complexity and improve energy efficiency,we use an improved K-means-based user clustering method for millimeter wave downlink massive MIMO-NOMA system to achieve virtual MIMO pairing of uplink users.In the millimeter wave downlink massive MIMO-NOMA system,we propose an improved K-means algorithm based on the millimeter wave user channel correlation and channel gain variability.Then,we design a low-complexity uplink virtual MIMO user pairing algorithm based on downlink user clustering because of the strong correlation of user channels in the same cluster and low inter-cluster interference.Simulation results show that the proposed joint uplink and downlink user clustering algorithm enables mobile terminals in uplink sub-6GHz and downlink millimeter wave systems to achieve higher energy efficiency and lower latency during task migration. |
PDF Full Text Request