Research On Massive Random Access For 5G Radio Transmission

As two main driving forces for the development of mobile communications in the future,Mobile Internet and the Internet of Things(Io T)provide a wider application prospect for 5G.However,due to the diverse wireless connection requirements of terminal equipment and the explosive growth of data traffic,the wireless technologies design for 5G faces many challenges.For instance,the access scenarios in 5G will be more complex than all the four previous generations of mobile communications.And the existing access schemes can not support the simultaneous connection of large-scale terminal equipment with diverse requirements.Based on this background,this dissertation investigates novel random access schemes suitable for 5G relative scenarios and performance requirements.The relevant research contents are summarized as follows:Firstly,a novel random access scheme based on Analog Bloom Filter(ABF)is developed.In order to reduce the computational complexity brought by multiple hash operations during ABF signature generation,we propose a single-hash ABF signature generation algorithm which includes a hash operation and several bit operations.Based on this,we apply the idea of approximate message passing to joint device activity detection and channel estimation with ABF signature.The simulation results show that the accurate device activity detection as well as channel estimation can be obtained in our proposed scheme.More specifically,our proposed method can support more active devices with high reliability.In addition,we propose a zone strategy,which changes the ABF signature generation zone according to user's angular coordinate,to further improve the system performance and give some analysis.Secondly,a power domain Non-orthogonal multiple access(PD-NOMA)which is suitable for uplink uncoordinated transmission is proposed.Different from conventional PD-NOMA scheme which often requires coordination mechanism,we use a set of pre-defined power levels and random power selection to achieve up link multi-user non-orthogonal transmission.Based on this,we apply the proposed scheme to two types of random access,while the corresponding schemes are called grant-free PD-NOMA-RA scheme and grant-based PD-NOMA-RA scheme,respectively.In order to evaluate the performance of the proposed two schemes,a unified multiaccess analysis model is first proposed.In this model,channel has three states and the state transition is modeled as a discrete-time Markov renewal process.Furthermore,through taking the overhead into account,the effective throughput of the two schemes are compared and analyzed.Simulation results show that both schemes can significantly improve system throughput without any bandwidth expansion and support more users.Finally,we put forward a novel random access scheme based on uplink code domain Nonorthogonal multiple access(CD-NOMA).Different from the conventional CD-NOMA which requires an access point(AP)allocating the specific pattern for each user before it can access the channel,no patterns will be pre-allocated in the proposed scheme and all users make fair contention to all unoccupied patterns,as does in CSMA/CA(Carrier Sensing Multiple Access with Collision Avoidance),a well-known random access scheme used in Wi Fi.On the user side,a pattern detection factor graph is introduced with pattern state node,and the pattern state is detected based on the soft information transition on the graph.On the AP side,the multi-user superimposed signal is recovered using a factor graph-based BP algorithm to achieve multi-user non-orthogonal transmission.Based on this,the NOMA/CA protocol with backoff procedure is further designed and the analysis of saturated throughput was given with some proper modification.The simulation results vertify that the uplink saturated throughput can be greatly improved and more channel connections can be supported.