Research On Pilot Allocation In Distributed Antenna Systems

In recent years,the rapid development of mobile communication services promotes the research of the fifth generation(5G)mobile communication system.Large-scale distributed antenna system(DAS),as a promising antenna deployment architecture in 5G system,possesses much higher macro-diversity gain and smaller transmission distance.In addition,it can improve spectral efficiency,system coverage and energy efficiency.Uplink pilot training is a common method for estimating channel state information(CSI).However,due to the limited pilot resources and huge number of users in the large-scale DAS,pilots have to be reused by users.Thus,pilot contamination(PC),which seriously affects the system performance,will occur in the pilot training phase.In this paper,the closed expression of downlink reachable rate in DAS is derived under pilot contamination.What is more,in order to reduce the influence of pilot contamination on system performance,two pilot scheduling algorithms based on game theory and mutual pollution index are proposed.Firstly,the statistical model of wireless channel,including path loss model,shadow fading model and small-scale fading model,is introduced.In order to facilitate the later research,two classical channel estimation algorithms,which are called least square(LS)algorithm and minimum mean square error(MMSE)algorithm,are elaborated.In addition,some common pilot scheduling algorithms are briefly described.Besides,maximum ratio transmission(MRT)precoding and zero forcing(ZF)precoding are introduced to prepare for the research of downlink achievable rate.Then,the closed expression of downlink reachable rate in distributed antenna system is studied and derived.In the case of pilot contamination,two different methods are used to derive the downlink reachable rate under MRT beamforming and ZF beamforming respectively.Based on Jensen inequality and Gamma distribution,the closed expression of downlink reachable rate under MRT beamforming is derived.In addition,according to the Gamma approximation of non-isotropic channel vectors and the properties of Gamma distribution,the reachable rate under ZF beamforming is derived.The accuracy of the closed expressions are verified by Monte Carlo simulation.Moreover,the simulation results show that the pilot contamination has a great loss on the system performance.Next,in order to reduce pilot contamination,a pilot scheduling algorithm by user clustering is proposed.In this paper,the influence of pilot contamination on the system performance is clarified in the system model.The problem of reducing pilot contamination is transformed into the optimization problem of minimizing the mean square error of channel estimation.According to the knowledge of game theory,the preference function,the principle of user clustering and the effective conditions for the stability of cluster structure in user clustering scenarios are determined.In addition,a heuristic pilot scheduling algorithm based on game theory is proposed.The simulation results show that the performance of the proposed pilot scheduling algorithm is close to that of the exhaustive algorithm.Finally,the pilot scheduling algorithm is deeply studied and a heuristic pilot scheduling algorithm based on mutual contamination index is proposed.Under the two scenarios of whether the channel model considers the angle of arrival(AoA),the concept and design process of mutual pollution index are defined and introduced respectively.The problem of reducing the mean square error of channel estimation is transformed to an optimization problem related to mutual pollution index.What is more,a pilot allocation algorithm based on mutual pollution index is proposed,which is applicable to both channel models.In addition,the proposed algorithm is compared with existing algorithms and the complexity of different algorithms is analyzed as well.Through analysis and comparison from different perspectives,it is proved that the algorithm based on mutual pollution index has the characteristics of high performance and low complexity.