Resource Allocation In Wireless Networks Based On Non-orthogonal Multiple Access

The prosperity and development of mobile communication and Internet technology promote the emergence of a large number of innovative applications,resulting in a large number of intelligent terminal devices,which are widely used in industrial Internet of Things,smart power grid and other scenarios.However,computing resources and power supply on these devices are limited.It is difficult to meet the processing requirements of computation intensive and latency sensitive applications.The demand of large-scale data transmission intensifies the link transmission burden of traditional wireless network,and the scarce orthogonal time-frequency communication resources cannot support the access of massive devices.Non-orthogonal multiple access(NOMA)uses the same time-frequency resource block to transmit multiple data streams,which can be multiplexed in the power domain.The receiver uses the successive interference cancellation(SIC)technology to complete the superposed signal decoding according to the multi-channel difference.Therefore,non-orthogonal multiple access technology can support large-scale connection,improve spectral efficiency,and become one of the key access technologies in the future communication network.NOMA technology artificially introduces interference between users.The channel conditions between multiple users are different,the quality of service requirements are different,and the wireless environment is complicated and changeable due to user mobility.Therefore,the wireless network resources must be rationally allocated to achieve performance optimization.In this paper,the optimal resource allocation strategy of non-orthogonal multiple access technology in a variety of wireless network architectures and communication network technologies is studied,mainly from the following three scenarios:(1)Partial offloading and resource allocation strategies for single user and multiple base stations.Mobile edge computing(MEC)technology is used to improve the computing power of network.MEC deplores powerful computing and storage resources at the edge of the access network to allow users to migrate local computing tasks to the edge base station for processing,providing efficient,fast and energy saving computing services.Consider the case of intensive deployment of edge base stations,where one user can connect to multiple edge base stations.The user divides the computing task into several sub-tasks,and uses NOMA to broadcast all the sub-tasks to several edge base stations,and each base station decodes its own target sub-task and performs the computing process.In order to minimize the energy consumption of all users,a resource allocation algorithm based on alternating direction multiplier method is proposed to jointly optimize task segmentation,transmission time and local CPU frequency.The convergence and energy saving advantages of the algorithm are verified by simulation.(2)Partial offloading and resource allocation strategy of multi-user and single base station.In the traditional cell network,the central base station provides communication computing service for all users in the cell.The computing tasks generated by the user are treated as a whole and can be left on the local device for computation or offloaded to the base station for remote processing via the NOMA mode.Considering the randomness generated by the task,the uncertainty of wireless channel,and the dynamic nature of network resources,the equipment offloading decision,transmission power,and CPU frequency resources were jointly optimized to maximize the long-term network utility(defined as a convex function of weighted data rate),while satisfying the network stability and user fairness.Using Lyapunov optimization technique,the long-term stochastic optimization problem is transformed into a single time slot deterministic optimization problem.By introducing auxiliary variables,the non-convex optimization problem is transformed into convex optimization problem,and an online resource allocation algorithm is designed.The asymptotic optimality of the algorithm with all or part of the device state information and the tradeoff between network utility and stability are theoretically analyzed.Simulation results show that the proposed algorithm has higher network throughput and fairness than the benchmark algorithm.(3)In order to further improve the network capacity and transmission reliability,large-scale antenna array is configured at both the transmitter and receiver,and the boundary of the traditional cell is removed,so that a user can connect to multiple base stations,forming a massive multi-input multi-output(MIMO)network.Non-orthogonal multiple access is used in two stages.First,the channel estimation process adopts non-orthogonal pilot sequences.Considering imperfect SIC decoding,statistical channel characteristics and pilot pollution,the closed-form expression of signal-to-noise ratio is derived.The pilot sequence length and antenna array size are analyzed step by step.The influence of pilot length on the accuracy of channel estimation is analyzed by simulation.Second,NOMA mode is adopted in the data transmission stage.Aiming at maximizing the network energy benefit(the ratio of throughput to energy consumption),Dinkelbach method and fractional programming are used to design the uplink power control algorithm.Simulation results show that the cell-free network can support larger scale access,and the proposed algorithm has better energy efficiency than benchmarks.