Research On The Superimposed Pilot Scheme In Massive MIMO System

As a one of the key technologies for the fifth generation of mobile communication technology,massive Multiple-Input Multiple-Output(MIMO)technology has attracted the attention of experts and scholars because of its great performance advantages.Benefit from a large number of antennas in Base Stations(BS),channels from different users to the BS becomes asymptotically orthogonal,and hence,significant improvement in spectral efficiency and energy efficiency can be achieved by using simple linear processing,such as Zero-Forcing(ZF)or Minimum Mean Square Error(MMSE).However,the massive MIMO system can achieve good performance heavily depends on condition that the BS can access to accurate Channel State Information(CSI).Thus,channel estimation becomes very important issue in the system above.Usually,the channel estimation is implemented based on pilot system.There are two conventional ways to use the pilots.The first one is time multiplexing scheme,that is,user spends a period of time transmitting a pre-designed pilot signal to the BS firstly,and then,the BS uses the received signals to estimate the CSI.In order to obtain an acceptable accuracy of the channel estimation,it has been shown that the length of the pilot signal should be greater or equal to the number of users in the system.Otherwise,the accuracy of the estimation will be severely reduced,and so-called “pilot contamination" occurs,which severely degrades the performance of the system.Once the length of the pilot signal satisfies the condition above,considerable time resource will be consumed since no data is transmitted in the pilot transmission period if the number of users is very large.The second way to acquire the CSI is the superimposed pilot scheme.The basic idea of this scheme is to simultaneously transmit the pilot signal and the data signal together with appropriate power,and the BS estimate the CSI by viewing the data signal as noise.Compared with the time multiplexing scheme,the superimposed pilot scheme saves the time dedicated to transmitting the pilot signal,but also brings mutual interference between the pilot signal and the data signal in the channel estimation and data decoding process.Based on multi-cell multi-user massive MIMO system uplink channel,this paper studies and improves the superimposed pilot scheme.The specific work includes the following two aspects.On the one hand,for the slow-varying channel,based on the basic idea of projecting the data signal and the pilot signal into different spaces as much as possible,a superimposed pilot scheme based on spatial multiplexing is proposed.The scheme can not only eliminate partial interference during channel estimation and data decoding,but also mitigate the influence of pilot contamination,thereby improving system performance.Furthermore,for the power allocation between the data signal and the pilot signal,the optimal solution is given in the manner of mathematical derivation.Finally,an iterative algorithm that can improve the initial channel estimation is designed by using the feature that some data signals are decoded with less interference.On the other hand,for the characteristic of small channel coherence time in fast-varying channel,a superimposed pilot scheme is redesigned,and the corresponding MMSE-based channel estimation and optimal power allocation scheme are also given.Finally,the superiority of the proposed schemes are verified by simulations.