Research On Asynchronous Reception In Cell-free Massive MIMO Systems

Traditional cellular networks face some problems such as large signal to noise ratio change and severe inter-cell interference.Compared with cellular networks,cell-free networks have greater advantages in spectral efficiency,energy efficiency,and coverage probability.In cellfree networks,each user simultaneously receives signals sent from all access points(APs).Due to the different distances between APs and users,there is a delay in the arrival of signals.In cell-free massive multiple input multiple output(MIMO)systems,multiple APs serve multiple users and it is difficult to coordinate the overall situation.The asynchronous reception effect will inevitably occur,which will seriously affect the performance of the system.This thesis mainly analyzes the influence of the asynchronous reception effect on the performance of a cell-free massive MIMO system,and designs a distributed intelligent algorithm to reduce the influence of the asynchronous reception effect on the system performance.The main work of this paper is as follows.Firstly,a cell-free massive MIMO downlink transmission model and channel estimation mechanism are analyzed.The precoding schemes commonly used in communication systems are expounded,and their respective advantages and disadvantages are analyzed and compared.Moreover,the closed expressions of the user achievable rate of the system when different precodings are used are given.In addition,reinforcement learning theory is also expounded,including Markov decision process(MDP)and its common solution methods,and the analysis of multi-agent deep deterministic policy gradient(MADDPG)is emphasized.Then,the impacts of asynchronous reception effect on the performance of a cell-free massive MIMO system are analyzed.We first derive the minimum mean squared error(MMSE)channel estimation under the influence of asynchronous reception effects,which reveals that asynchronous reception effects introduce a phase shift matrix to the estimated channel state information(CSI).This reduces the quality of the channel estimation and increases the difficulty of deriving the closed expression of the user achievable rate.Given the imperfect CSI,we derive closed form expressions of user achievable rates in asynchronous reception systems with maximum ratio transmission(MRT)and regularized zero-forcing(RZF)precoders.The simulation results verify the accuracy of the closed expression and the effect of asynchronous reception will significantly reduce the spectral efficiency of the system.In addition,the sum rate of the asynchronous receiving system is inversely proportional to the subcarrier sequence number.The system performance degrades more significantly with the increase of the signal-to-noise ratio or the number of antennas.Finally,based on the analysis of the influence of asynchronous reception effect on the performance of cell-free massive MIMO systems,a distributed intelligent algorithm to reduce the asynchronous reception effect is studied.First,an optimization problem is established with the goal of maximizing the sum rate and the time adjustment interval as the constraint.And based on the MDP,the agents,observations,actions and rewards of the problem are designed.In order to achieve the goal of reducing the impact of asynchronous reception on system performance,a distributed intelligent algorithm based on MADDPG learning framework is designed to adjust the transmission time of base stations for base stations can cooperate with each other and balance the delay difference.The simulation results show that the proposed distributed intelligent algorithm can significantly reduce the influence of asynchronous reception effect on the spectral efficiency of the cellular-free massive MIMO system.Meanwhile,the performance of the distributed intelligent algorithm and the block compensation algorithm are compared through simulation.The effectiveness of the distributed intelligent algorithm is verified.