Study On User Clustering And Power Allocation Algorithms In MIMO-NOMA System

The rapid development of mobile internet and internet of things(IoT)had brought people a more convenient life,but also posed a greater challenge to 5G.In order to cope with the problems caused by massive traffic and the lack of frequency band resources,key technologies for higher spectrum efficiency are crucial.As the core technologies of both 5G and future mobile communication,power domain non-orthogonal multiple access(NOMA)and multiple-input multiple-output(MIMO)technologies can provide larger spectral efficiency without adding additional system bandwidth by power domain reuse and air-space multiplexing.Based on this,the combination of NOMA and MIMO technologies can exponentially increase spectrum efficiency,achieve massive multi-terminal access and realize large-capacity wireless transmission.In MIMO-NOMA systems,multiple users perform power domain non-orthogonal superposition transmission on the same time-frequency resource block,which causes a strong inter-user interference problem.At the same time,receivers of MIMO-NOMA systems demodulate through serial interference cancellation(SIC)receiver,in which the demodulation order depends on the transmit power of the receiving users.Therefore,how to allocate power so that both the receiver's decision condition and the user's service quality can be guaranteed still require further research.Based on these above,this thesis focuses on the key technologies in MIMO-NOMA systems,and the main research contents are as follows:(1)The signal transmission and processing process in MIMO-NOMA systems are studied,and signal transmission differences analysis between uplink and downlink in MIMO-NOMA systems are the key point.By analyzing the signal transmission process in uplink and downlink respectively,the beamforming matrix design and detection vector design which affect the uplink and downlink systems performance as well as the main points of the transmit power control technology in the uplink system are deduced.(2)The user clustering algorithm in MIMO-NOMA downlink systems is studied.Firstly,the classic user clustering algorithms are introduced,their advantages and disadvantages are analyzed through simulation.Secondly,the Jain fairness index is introduced to measure the user fairness in the systems,and an improved dynamic clustering method for MIMO-NOMA users is proposed with the goal of ensuring user fairness.Simulations indicate that the proposed algorithm can further improve the throughput performance of systems under the premise of ensuring user fairness.(3)The user power allocation algorithm in MIMO-NOMA downlink systems is studied.Firstly,the advantages and disadvantages of each classic user power allocation algorithm are analyzed through simulation experiments.Secondly,the user power allocation algorithm of users in the same cluster is studied with the goal of maximizing the minimum user SINR.The non-convex optimization is transformed into a quasi-convex optimization problem,and a simple dichotomy algorithm is proposed to obtain a suboptimal solution for the optimization goal.In this process,the conditions that the power distribution factor of each user in systems needs to meet under acceptable QoS are derived in a recursive manner,which ensure the fairness of users and narrow the search range of the proposed optimization algorithm.Simulations indicate that,compared with MIMO-OMA systems and fixed power allocation algorithm,the proposed algorithm can effectively improve user fairness under the premise of guaranteeing user QoS.