A Study Of NOMA Network Based On Buffer-aided Relaying

Non-orthogonal multiple access(NOMA)technique based on power domain is considered as one of candidate multiple access schemes of 5G network.It can transmit signals of multiple users in the same time-frequency resource and improve the spectrum efficiency of the system significantly.By combining buffer technique with relay technique,the buffer-aided relaying technique changes the original fixed transmission mode of the relay in which the relay receives signal in the previous time slot and sends signal in the next time slot,so that the system can better utilize the channel state information and obviously improve the performance of relay cooperative user.The combination of NOMA technique and buffer-aided relaying technique can effectively improve the performance of cell edge users and the transmission reliability of the system.This thesis studies a relay cooperative NOMA network,in which base station directly communicates with the near user while communicating with the far user through a relay.According to the traditional NOMA power allocation strategy,the base station allocates more power to the far user,so that its throughput tends to reach an upper limit as the signal to noise ratio(SNR)increases when the relay decodes the far user's signal according to the successive interference cancellation technique.At the same time,the near user's throughput can continue to increase with increasing SNR,which makes although NOMA technique brings the fairness of user access,it cannot guarantee the fairness of user throughput.In order to solve this problem,this thesis applies buffer-aided relaying technique to the network and proposes three link selection strategies.Firstly,a link selection strategy based on throughput is proposed.In order to maximize the far user's throughput,an optimization problem is constructed and the optimal link transmission sequence is obtained by using the Lagrangian multiplier method.Simulation results show that the far user's throughput can be significantly improved and the fairness of throughput can become better when the SNR is relatively high.Then,considering the data delay caused by the buffer-aided relaying technique,this thesis proposes a link selection strategy based on the buffer state.The time for data to stay in the buffer is controlled by introducing a buffer threshold and the entire transmission process is modeled by the Markov chain model.Simulation results show that the far user's interruption probability under this strategy can be significantly reduced.Finally,considering that the above two strategies still cannot change the situation that the far user's throughput will reach a fixed value when the SNR is relatively high,this thesis also proposes a link selection strategy based on power allocation.In addition to using buffer-aided relaying technique,this strategy also proposes a new power allocation strategy,that is,no longer allocating low power to near user,high power to far user,but in turn,allocating low power to far user,high power to near user.In addition,considering the high computational complexity when using the optimal decision function in the proposed strategy,this thesis also studies the case of a sub-optimal decision function and gives the expression of the two user's throughput.Simulation results show that,under this proposed strategy,the system throughput is improved when the SNR is relatively low and the fairness of throughput between users is significantly improved when the SNR is relatively high.