Research On User Clustering And Energy Efficient Power Allocation Of Millimeter Wave Massive MIMO-NOMA System

The combination of mm Wave massive multiple input multiple output(MIMO)architecture and Non-Orthogonal Multiple Access(NOMA)technology is considered as a solution to meet the ever-increasing data rate demands.Consider the downlink of a mm Wave MIMO-NOMA system,where the base station(BS)is equipped with a massive antenna array.The antenna array uses hybrid precoding to reduce the number of radio frequency(RF)chains required,reducing performance loss and lowering hardware costs.In the millimeter-wave massive MIMO-NOMA system,the inter-user interference has a significant impact on the system performance,and a reasonable user clustering scheme and power allocation algorithm are of great significance.This thesis proposes a user clustering strategy,hybrid Precoding schemes and power allocation optimization algorithms to ensure improved spectral efficiency and maximize system energy efficiency.The main research contents of this thesis are as follows:First of all,in order to improve the system performance,the influencing factors of channel gain and channel correlation need to be considered at the same time.We select the user with the largest channel gain as the first initial cluster head,introduce a threshold that can measure the channel correlation,and continuously self-autonomous through the threshold.Adapt the update and channel correlation to determine the cluster heads of other user clusters.In the later stage of the algorithm in this paper,the remaining users can be iteratively assigned to different clusters,so as to obtain more accurate clustering results and user scheduling transmission after clustering.In this way,the spectral efficiency of the system can be effectively improved.Secondly,due to the continuous increase in the number of RF chains required by the system,the energy consumption of the MIMO-NOMA system will increase significantly.We adopt a two-stage design method of a high-dimensional analog precoder and a low-dimensional digital precoder to significantly reduce the number of RF chains required,thereby reducing system energy consumption without significant performance penalty.The analog precoding is designed for all beams according to the previously selected cluster heads.After that,the users are clustered according to the correlation of the equivalent channels of the users.Finally,digital precoding is designed by selecting the user with the strongest equivalent channel gain in each beam.Compared with other linear precoding,low-complexity zero-forcing precoding can completely eliminate inter-cluster interference and cluster interference.It can completely eliminate the inter-cluster interference and the interference between data streams in the cluster,and it is also suitable for multi-antenna systems and multi-users.Finally,the energy efficiency maximization problem under the constraints of data rate and power is established.The power distribution optimization problem is a nonlinear optimization problem.To solve the constrained optimization problem,the general suboptimal solution algorithm is not necessarily the fastest.Its solution is not guaranteed to be the global optimal solution,and it is easy to obtain only the local optimal solution.It is difficult to directly solve the optimization problem because weak users are interfered by strong users and other clusters.Therefore,a swarm intelligence Whale optimization algorithm(WOA)algorithm is used to solve the optimal solution,random or optimal search agents are used to simulate the hunting behavior of humpback whales,and a spiral bubble net is used to prey and attack mechanisms.But there are some problems with this algorithm.When it reaches the late stage of the iteration,the convergence speed of the WOA algorithm will gradually slow down,and it may fall into a local optimum.In order to solve this problem,in the spiral predation stage,the weight factor and fitness value update are introduced to optimize and improve.In the process of solving the optimal solution of power distribution,the improved algorithm improves the global optimization ability,reduces the number of iterative solutions,and improves the search efficiency in subsequent iterations,and finally effectively improves the energy efficiency of the system.In this thesis,a reasonable user clustering algorithm,a two-stage hybrid precoding scheme and an energy-efficient power allocation optimization algorithm are designed in the system.Combined with the simulation results of different algorithms,the spectral efficiency and energy efficiency of the system have been greatly improved.