Research On Receiving Method Of Uplink Non-Orthogonal Multiple Access For Small Data Packets

With the popularization of smart terminals and the development of the mobile Internet,a large number of applications such as QQ,WeChat,Weibo,and other instant messaging services and machine to machine(M2M)communication services have generated a great number of small data packets which are sparse in time domain.Meantime,the number of mobile terminals and Internet of Things(IoT)devices has also shown a sharp rise,resulting in an increasing proportion of small data packets in communications networks.In face of the massive small packet services,the current Orthogonal Multiple Access(OMA)scheme faces the problems of low spectrum efficiency,limited number of connections,and heavy signaling overhead.Thus more efficient multiple access technologies have attracted widespread attentions.The NOMA scheme have significant advantages in terms of spectral efficiency and large connectivity support.For the problem of heavy signaling overhead,the uplink grant-free transmission mechanism needs to be further considered,i.e.,the users can randomly transmit data without the base station performing scheduling.For uplink grant-free NOMA systems of small packets,non-orthogonal allocation of resources can lead to greater interference among users,thereby increasing the complexity of the receiver.On the other hand,the process of scheduling is eliminated,thus the user activity is unknown in advance,the detection of active user and the demodulation of signals need to be jointly implemented at the receiver,it further increases the difficulty of designing the receiver.In response to these challenges,this paper will study how to make full use of the characteristics of small data packets to achieve the accurate user detection and the reliable signal demodulation.First,we analyze the uplink NOMA model of small data packets.Utilizing the user behavior sparseness naturally present in mass connections,the signal demodulation can be transformed into a typical block sparse recovery problem.And the current block sparse recovery algorithms are also summarized and compared.By analyzing the difference between small data packets and general sparse recovery,we apply the interference cancellation to block sparse recovery and generalize the basic steps of small packet demodulation.Then,in order to solve the problems that demodulation algorithm cannot be stopped in time and the users are likely miss or false detected,we studied the variation law of the residual energy and obtained an appropriate iterative termination conditions,which makes it possible to stop iterations in time when all active users have been detected.This method achieves an indirect estimation of the number of active users.Thus,reliable demodulation of block sparse signals can be achieved without relying on the number of active users(sparsity level)in advance.From the simulation results,the proposed method can accurately estimate the sparsity level in both general block sparse recovery and small packet demodulation.Finally,in order to better control the interference of multi-user demodulation,we studied the specific demodulation process,obtained the law of user detection order and the relationship between the detected users and decoding performance.Under a reasonable control of false detection,the decoding user group is suitably determined for each iteration.Thus a block sparse recovery algorithm based on successive group decoder(SGD)is proposed,to speed up the iterative process and improve the demodulation accuracy.At the same time,we analyzed the theoretical demodulation performance and computation complexity.From the simulation results,we can see that compared with current algorithms,the proposed method has obvious advantages in terms of the accuracy and complexity,the iterative process can converge faster,and the demodulated error frame rate is also lower.