Pilot Optimization Strategy For Cell-free Massive MIMO

The cell-free massive MIMO system eliminates the concept of "cell",and based on the distributed antenna system,it deploying a large number of distributed access points(AP),the distance between the AP and the user is shortened,the path loss is reduced,and a greater macrodiversity gain is obtained.However,too many users or too short coherence interval duration will cause users to reuse pilots,which will cause serious pilot contamination and affect system performance.In this thesis,the pilot optimization strategies in cell-free massive MIMO are studied.A pilot allocation algorithm based on iteration removal is proposed to solve the problem of pilot contamination in cell free massive MIMO systems.First,use the location information of each user to calculate the relative distance between users,establish a pilot-permit-circle with an appropriate radius for users who have not been assigned a pilot,and select this user pilotpermit-circle according to the relative distance between other users and this user.Users in the pilot-permit-circle are assigned mutually orthogonal pilots,and then users who have been assigned pilots will be removed from the user set and will no longer participate in the subsequent user selection and pilot allocation processes in the pilot-permit-circle.Finally,the greedy method of minimizing pilot contamination is used to updated the pilot for the user with the lowest rate.The proposed pilot allocation algorithm based on iteration removal prevents users who are close to each other from allocating the same pilot to a certain extent.Removing the users who allocate pilot from the user set can reduce the complexity of the algorithm.Finally,the performance of the users with the lowest rate can be guaranteed by the greedy method.The experimental results show that the pilot allocation algorithm based on iteration removal can effectively reduce the pilot contamination and improve the network throughput of cell-free massive MIMO system under the maximum pilot-permit-circle radius.Compared with random pilot assignment and greedy pilot assignment,the pilot assignment algorithm based on iteration removal improves the 5% outage rate of per-user uplink throughput by 23% and 11%,respectively.In order to reduce backhaul overhead and pilot inter-user interference,an AP selection algorithm based on the maximum desired signal is proposed.Firstly,the contribution rate of all APs to the user's desired signal is calculated and sorted in descending order,the minimum number of AP satisfying its cumulative threshold is selected as the effective service AP subset of this user.In order to prevent users with good channel state from affecting users with poor channel state,the minimum-maximum power control algorithm is used to assign pilot power control coefficients to users.In order to minimize the maximum normalized mean square error of all users,a non-convex optimization problem is established,which is transformed into a convex optimization problem by using Lagrange duality and first-order Taylor approximation.The results show that the proposed AP selection algorithm based on the maximum desired signal has a lower number of users accessing APs,and the power control algorithm used has a faster convergence speed.The joint AP selection and pilot power control scheme is reducing the backhaul link overhead and improve the quality of service(Qo S)of users.In the case of considering the pilot power control algorithm,compared with the full connection and the AP selection algorithm based on the largest large-scale fading coefficient,the AP selection algorithm based on the maximum desired signal improves the 5% outage rate of per-user uplink throughput by 15% and 9%,respectively.