Research On User Clustering And Power Allocation In NOMA System

The development of wireless communication technology has brought great convenience to human social life,and everyone’s social activities carry out information transmission at any time,and the efficiency and quality of personal information transmission should be a problem that needs to be studied in focus.Especially in 5G,6G and other high frequency communication technologies,there will be more terminals to access the Internet,and the data to be transmitted will be unimaginable before.Therefore,for the whole process of information transmission of wireless communication system,the research of each stage must be carried out objectively and deeply in order to provide a better guarantee for human life.In Non-Orthogonal Multiple Access(NOMA)system,to transmit the user’s data accurately,then the user must first be distinguished,that is,the user clustering,so that the base station can accurately identify each user.At the receiving end when users receive the information,they can accurately demodulate and get the information they want,which is the first step.After the user clustering,the issue to be considered is the different power allocation to different users according to the communication rules,which is also the key basis for the demodulation of user information.Therefore,the clustering and power allocation of users are the key issues to be studied.In this paper,these two issues will also be studied.The main research of the paper is as follows.First,the data clustering algorithm is studied,and a user clustering algorithm based on unsupervised learning is proposed.An optimization model for maximizing system capacity is established for the communication system model,and a new user clustering algorithm is proposed using the K-Means clustering algorithm and the features of the improved algorithm to reasonably classify users with respect to their characteristics.In a multi-user environment,through simulation analysis,it is found that the algorithm has significantly lower complexity and faster convergence than the previously used exhaustive search algorithm,and the system capacity is improved.Second,for the user power allocation problem,thanks to the iteration of evolutionary algorithms,the characteristics of updating for optimization,this paper compares the genetic algorithm and the particle swarm optimization algorithm,and finds that the particle swarm optimization algorithm works better.So this paper proposes to use the particle swarm optimization algorithm to allocate the user power.After simulation analysis,it is found that the allocation algorithm used in this paper is faster in reaching the optimal value,can effectively improve the system capacity and has higher stability compared with the existing algorithms.