Research On Key Technologies Of Wireless Communication For Large-scale Connection Scenarios

In recent years,with the rapid development of mobile Internet,the number of mobile devices and mobile data traffic show explosive growth.Facing with the requirements of Internet-of-Things(IoT)and ultra dense network deployments,the future cellular mobile communication needs not only higher network throughput,but also improved capability to support massive connectivity.Based on the requirements of the upcoming 5th Generation(5G)mobile communications,this dissertation discusses some promising techniques to provide higher throughput and more connectivity,such as Non-Orthogonal Multiple Access(NOMA),grant-free data transmission protocols,heterogeneous cellular networks(HetNets),Device-to-Device(D2D)communications,and wireless caching networks.Then this dissertation propose a series of effective solutions to the challenges of massive connectivity scenarios from both Physical(PHY)layer and Media Access Control(MAC)layer aspects.Considering the PHY and MAC technology to support the massive connectivity of 5G wireless networks,a lot of research in both academia and industry has been conducted these years.Although the Release-15 version of 5G standard has been officially published by the 3rd Generation Partnership Project in June 2018,the debate on the multiple access technology,the data transmission protocol and the network architecture adopted by massive connectivity scenarios is still fierce.In order to effectively apply the newly proposed key technologies for 5G in the massive connectivity scenarios,this dissertation improves the protocol design and conduct comprehensive analysis,comparison,and evaluation of these key technologies.The contents and innovations of this dissertation are summarized as follows:1)Most of the existing research on grant-free non-orthogonal multiple access(GFNOMA)requires the prior knownledge of the number of active users,which is not a practical assumption in actual wireless communication systems.Only rough estimation of the active users can be obtained,which leads to serious missed and false detection and increases the complexity of the joint user detection and data decoding after the MUD.Therefore,this dissertation proposes a novel receiver for uplink grant-free Sparse Code Multiple Access(SCMA)that relies on no prior knownledge of the number of active users.The Detection-based Group Orthogonal Matching Pursuit(DGOMP)MUD algorithm is proposed to obtain accurate active user detection.The Joint Mass Passing Algorithm(JMPA)is modified by taking the channel gain and noise power into consideration when calculating the prior information of the zero codeword.Moreover,this dissertation analyzes the success detection and channel estimation error performances of GF-NOMA systems via stochastic geometry.2)When Sparce Code Multiple Access(SCMA)was proposed,all research in both academia and industry were on the theritcal analysis and optimization of SCMA systems.However,no research was on the hardware implementation of SCMA systems,which slowed down the commercialization and standardization progress of SCMA.Therefore,we study the implementation of SCMA systems with Field Programmable Gate Array(FPGA)platforms and the problem of fixed-point optimization design.This dissertation describes the implementation and optimization of SCMA systems on fixed-point platforms,and analyzes the complexity of SCMA decoding algorithms.To reduce the processing delay of SCMA decoder,we carefully design the Verilog HDL-based log-MPA module,and give the modular design supporting parallel structure.The bit error rate performance of fixed-point log-MPA is about 0.3 dB worse than that of float-point logMPA.The block error rate performance of hardware experiment is about 0.22 dB worse than that of software simulation.3)From the aspect of network architecture to support massive connectivity,although there has been a lot of study on D2D-enabled HetNets,there is no study on applying NOMA technique to further improve the network throughput.The existing study on the resource management of NOMA systems focus either user pairing or access point selection,without consideration of the joint optimization of them.This dissertation proposes a novel framework on D2D-enabled HetNets with NOMA,where NOMA is invoked to serve more users simultaneously,and D2D-enabled multi-hop transmission is established to enhance the signal reception of cell-edge users.This dissertation proposes some effective strategies and realizable algorithms for resource scheduling and interference management.Our study in this area can be separated into two parts:Firstly,we consider the single-D2D link model and formulate two optimization problems with the maximum-minimum-rate and the maximum-sum-rate criteria,respectively.To effectively solve the problems,we develop a two-step method,which can always reach the closed-form solution.Secondly,we extend our system model to a multiple-D2 D link model and investigate joint power allocation and user scheduling.To overcome the challenges from the large amount of Channel State Information(CSI)measurements and feedbacks,a robust algorithm is proposed to optimize the ergodic user rate based on incomplete CSI,which yields 95% of the network throughput of the method with complete CSI.To obtain the solution to the joint user pairing and access point association problem,a low-complexity three-sided matching algorithm is proposed,which yields 80% of the network throughput of the exhaustive searching algorithm.4)Adopting the full duplex technique into NOMA systems can significantly improve the network throughput,which however has not been studied by both academia and industry when we began our research.In this dissertation,we focus on the performance analysis and optimization of Cooperative Full-duplex Relaying NOMA(CFR-NOMA)systems.In particular,we first derive the closed-form expressions of outage probability and ergodic sum rate with fixed power allocation.Then we analytically obtain the closed-form expressions fo the solutions to the outage probability optimization problem and the minimum user achievable rate optimization problem.5)This dissertation invokes the cooperative multi-point technique to mitigate the inter-link interference and improve network throught in D2D-enabled wireless caching networks.This dissertation studies the resource allocation aspect of the aforementioned networks and formulates a joint link scheduling and power allocation problem to maximize the overall throughput of cooperative D2 D links(CDLs)and non-cooperative D2 D links(NDLs),which is non-deterministic polynomial-hard.To solve the problem,we decompose it into two sub-problems,which maximize the sum rates of the CDLs and the NDLs,respectively.For CDL optimization,we propose a semi-orthogonal-based algorithm for joint user scheduling and power allocation.For NDL optimization,we propose a novel low-complexity algorithm to perform link scheduling and develop a Difference of Convex functions(D.C.)programming method to solve the non-convex power allocation problem.