Pilot Optimization Strategy For Cell Free Massive MIMO System

In order to meet the strict requirements of service quality for future mobile networks,network densification is essential.However,due to inter-cell interference,neither ultra-dense network nor massive multiple-input multiple-output(MIMO)technology can meet the growing data rate demands of sixth-generation wireless communications,and cell free massive MIMO system is born on demand.Cell free massive MIMO system uses densely deployed distributed access points(APs)to serve users.Through a large number of APs,cell free massive MIMO achieve channel hardening and favorable propagation conditions greatly reducing small-scale fading,noise,and interference from other users.While inheriting the advantages of traditional massive MIMO,the cell free network also suffers from the problems of limited pilot frequency resources and inter-user interference,which seriously impacts system performance.As a consequence,the pilot frequency optimization problem of cell free massive MIMO system is studied in this paper.In order to reduce the impact of pilot contamination on the cell free massive MIMO systems,the greedy pilot assignment based on location with pilot power control algorithm(GPABL with PPC)algorithm is proposed in this paper.In terms of pilot allocation,the greedy pilot assignment based on location(GPABL)algorithm is used to ensure that the pilots assigned to users who are close to each other are different,and then the pilot sequence of the user with the lowest downlink rate is updated to reduce pilot contamination in cell free massive MIMO systems.In terms of power control,the pilot power control(PPC)algorithm is used to design pilot power coefficients for all users,so that the maximum of the normalized mean square error of all users is minimized,thereby improving the channel estimation accuracy in the uplink training phase and maximizing system capacity.Then,the two pilot optimization methods are combined to suppress the impact of pilot contamination on the system to the certain extent.In addition,in order to mitigate the inter-user interference in cell free massive MIMO systems and reduce the backhaul link overhead of the system,this paper proposes a largest channel-estimation-coefficient-based AP selection algorithm(LCAS).Compared with the large-scale fading coefficient,the channel estimation coefficient can better reflect the change of the channel state information,and the channel estimation coefficient changes relatively slowly over time.Therefore,this algorithm selects APs based on the channel estimation coefficients between APs and users.Then,users are sorted according to their channel estimation coefficients,and AP subsets are assigned to users with poor channel quality first.From large to small,the minimum number of APs that can satisfy the cumulative threshold is selected as the effective service AP subset for users.When selecting AP subsets,in order to ensure that some APs are not overused,a usage threshold is set for each AP.If the channel gain effect of the AP selected by the user has exceeded the set minimum threshold,and the use frequency of the AP has reached the AP usage threshold,the user will no longer consider selecting this AP as the user’s AP subset;once no better choice is found in the subsequent AP options,the system will select this AP as the user’s AP subset.Finally,the network throughput of the cell free massive MIMO with only AP selection and joint AP selection and pilot power control was compared.The numerical results show that combining the two methods can improve the performance of cell free massive MIMO system by approximately 14%.