Research On Computation Offloading And Resource Allocation For Backscatter-aided And RF-powered MEC

The rapid development of mobile communication technology and the popularity of 5G communication have given rise to various new scenarios,such as virtual reality,autonomous driving,smart storage,and smart homes.Traditional Mobile Cloud Computing(MCC)aims to offload the complex computing tasks of mobile users to cloud servers in the core network to solve the problem of insufficient computing power of mobile users.However,due to the long distances between cloud servers and mobile users,multi-hop routing and forwarding,and massive data aggregation,large latency is generated,making it impossible to meet the low-latency requirements of these new scenarios.The emergence of Mobile Edge Computing(MEC)can effectively reduce the data transmission and processing latency of mobile users by deploying edge servers close to users.By offloading tasks to MEC,we can avoid long-distance transmission.Therefore,MEC can better meet the low-latency requirements of mobile users and then MEC is a key technology in 5G.To address the issue of limited battery capacity of mobile devices,RF-powered technology has emerged as a promising solution due to its ability to cater to the high endurance demand of users.However,the traditional Harvest-then-Transmit(HTT)technology in RF-powered suffers from a double proximity effect.Then,resulting in users located far from the RF source experience communication failure due to energy limitation.To overcome this challenge,backscatter communication has been proposed as a viable alternative.Unlike the HTT technology,backscatter communication enables mobile devices to communicate by reflecting RF signals from the surrounding environment,thus significantly reducing the communication energy consumption of users.In this thesis,users in MEC systems can offload tasks by utilizing HTT communication in conjunction with different backscatter communication techniques.The main contributions of this thesis include:This thesis considers a backscatter-assisted and RF-powered MEC network with multiple edge users(EUs),a hybrid access point(H-AP),and several carrier emitters(CEs).Each EU can choose to compute task locally or offload task to the MEC server based on its energy status,task size,etc.EUs can offload tasks using HTT communication,ambient backscatter(AB)communication,or bistatic backscatter(BB)communication.In addition,device-to-device(D2D)communication is introduced to allow EUs that are far away from the H-AP to complete task offloading with the help of other EUs through D2D communication.The thesis formulates a residual energy maximization problem for the considered system,and proposes the joint task processing/offloading mode selection and system resource-allocation schemes.We aim to maximize the residual energy of all energy harvesting based EUs while ensuring that EUs can complete their tasks,by jointly optimizing EUs’ task processing mode(local computing mode or edge computing mode),task offloading mode,and system resource-allocation.Since the formulated optimization problem is non-convex with multiple coupled variables,we adopt the block coordinate descent(BCD)method to decompose it into several subproblems.Then,we utilize the alternating directions method of multipliers(ADMM)to transform the non-convex subproblems into convex subproblems,and propose the distributed joint task processing/offloading mode selection and system resource-allocation schemes.Finally,we verify and evaluate the performance of our proposed scheme through numerical analyses,which show that our proposed scheme outperforms the other benchmark schemes in terms of the residual energy and the task completion rate of EUs by jointly selecting the task processing mode and the task offloading mode for each EU.