Research On Key Techniques Of Massive Random Access System For M2M Communications

The ever-evolving Internet of Things(Io T)has been increasingly reforming our everyday life with various applications,such as smart homes,smart farming,industrial automation,and intelligent transportation.To achieve the pervasive connectivity of Io T,a growing number of Io T devices will connect to the communication networks.Moreover,novel enabling technologies for massive access are becoming significant and emergent to accommodate the connectivity requirement of explosively growing Io T applications.The massive Io T is widely envisaged as one of the pivotal tendencies for the upcoming beyond5G(B5G)and 6G networks.Compared with conventional human-type communication(HTC),one of the key features of machine-type communication(MTC)in Io T is the massive access requests.The transmission of access requests and the acquisition of channel state information(CSI)are generally conducted by preamble(pilot)in the initial uplink access.However,due to the limited coherence interval,the number of available preambles in uplink random access is limited and it is not possible to assign the dedicated preamble to each active MTC device(MTDs).This mismatch between the preamble and active MTDs incurs severe preamble collisions,which significantly decreases the access performance.In addition,the peak throughputs of the existing access schemes are generally significantly limited by the number of preambles,which cannot well fulfill the requirement of massive access.To cope with these problems,the key techniques of massive random access system for machine-tomachine(M2M)communications are studied and analyzed from three aspects: random access in cellular networks,random access in distributed networks,and access control(optimization)in random access systems.The main research contents and contributions of this dissertation are listed as follows:1.In terms of grant-based random access in cellular networks,we propose a spatial filtering based random access(SFRA)protocol,in which the angle of arrival(AOA)information provided by massive MIMO(m MIMO)is utilized to distinguish colliding MTDs.To improve preamble utilization as possible,a group-based preamble reuse scheme is further proposed to eliminate the influence of multi-path angular spread on the detection of preambles at the base station(BS).The protocol can achieve the preamble reuse effectively and improve the access capacity significantly.Secondly,in terms of grant-free random access(GFRA)in cellular networks,we propose a grant-free random access protocol with channel differentiation(CD-GFRA)in m MIMO systems with coexisting line-of-sight(LOS)and non-LOS(NLOS)channels.In this scheme,MTDs with angle-dependent LOS channels access the network directly,while NLOS-channel MTDs transmit with the assistance of the random relay.Moreover,AOA-based channel estimation is further proposed to improve the access capacity of LOS-channel MTDs.2.In terms of grant-based random access in distributed m MIMO,we first propose a clusteraided collision resolution random access(CACRRA)protocol to mitigate severe pilot collisions.In this protocol,the difference in large-scale fading between different access points(APs)and MTDs is utilized to cluster APs dynamically in each time slot for each activated pilot.By leveraging the excellent properties of distributed m MIMO channels,i.e.,channel hardening and favorable propagation,the collision resolution is performed distributedly in each cluster,which enables effective pilot reuse at the same time.On this basis,a fixed cluster-based collision resolution random access(FCCRRA)protocol is further proposed,in which APs are grouped into fixed clusters based on geographic location information.Moreover,a modified precoded vector is adopted in the downlink precoded random access response(PRAR).Compared with the conventional strongest user collision resolution(SUCRe)protocols,FCCRRA significantly simplifies the processing procedure at the user equipments(UEs).The combination of fixed clustering collision resolution and modified PRAR significantly mitigates the complexity and energy consumption at both the CPU and UEs and greatly improves the access performance at the same time.3.In terms of access control in cellular networks,a spatial group based access class barring(SG-ACB)protocol is proposed to cope with the performance degradation of random access systems under overloaded scenarios.In this protocol,the timing advance(TA)information is utilized to assist the differentiation of colliding MTDs and preamble reuse.In addition,multiple ACB factors are used to provide precise access control.This scheme significantly reduces the average access delay and greatly improves the access capacity.In terms of access optimization in cell-free(CF)m MIMO networks,we first derive the theoretical result of the access success probability of MTDs for the available grant-free access transmission structure.Next,a novel transmission structure with grouped data superposition is proposed to further enhance the access performance.On the basis of deriving the access success probability and effective throughput of this novel structure,an optimization problem that maximizes effective throughput is further formulated to determine the optimal transmission strategy,which efficiently improves the access performance of grant-free access in CF m MIMO networks.