Research On AP Selection In Cell-free Massive MIMO System

At present,the new generation of information technology is accelerating its development and application,which is represented by artificial intelligence,quantum information and the Internet of Things and has more and more demanding requirements for the performance of mobile communication networks.Compared with traditional massive multiple input multiple output(MIMO)system,cell-free massive MIMO system can not only solve a lot of problems,such as frequent cellular switching,inter-cellular interference and uneven quality of service(Qo S),but also effectively suppress the interference between users and improve the quality of service by enhancing macro diversity.However,if each access point(AP)serves all users in the case of low load,the energy efficiency of the network system will be very low,which is not conducive to realizing green communication network and saving resources.Therefore,it is of great research value to select a part of APs for each user by AP selections and construct the user-centric cell-free massive MIMO system,so as to improve the energy efficiency of system.This paper presents two AP selection methods and the simulation shows that they can improve the energy efficiency:(1)This dissertation studies the downlink energy efficiency and proposes an AP selection method based on K-means clustering algorithm.K-means clustering algorithm is an unsupervised clustering algorithm commonly used in machine learning that can be used to group users.Firstly,the paper establishes the downlink energy efficiency maximization problem of the cell-free massive MIMO system.Then the AP selection method can group all users by K-means clustering algorithm and a part of APs are selected for each user group according to the large-scale fading factors.(2)This dissertation studies the uplink energy efficiency and proposes an AP selection method based on ZF receiver and sleep strategy,which is applicable to areas where the number of users fluctuates greatly.Firstly,this paper establishes the uplink energy efficiency maximization problem of the cell-free massive MIMO system with ZF receiver.This method has two different schemes when the number ratio of APs and users is different.On one hand,when the number radio is relatively large,this method uses the sleep strategy to temporarily sleep the unselected APs after selecting a part of APs for each user,in order to avoid unnecessary energy waste.On the other hand,When the number is small,this paper derives the expressions of user channel quality and channel quality between users and APs.This method sorts all users in descending order according to user channel quality,and connects each user with the AP with the best channel quality in turn.