Research On Decoding Method Of Low Complexity In Massive MIMO System

Wireless communication is one of the fastest growing technologies in communications systems.It has provided high-speed and reliable data transmissions.With the demands of speed and quality of data transmission increasing,demands of wireless communication technologies are also increasing.Massive MIMO technology is an emerging wireless transmission technology.With efficient spectral efficiency and excellent performance,it will become one of the next generation wireless communication technologies.The massive MIMO system refers to the base station using a large-scale antenna array.The number of antennas may be hundreds or more.It has advantages of the traditional multi-user MIMO system,and moreover,asymptotic orthogonality among channels of different users can improve the transmission rate of the system,increase the reliability of the link and reduce the interference.Although commonly used linear channel estimation and linear decoding methods,for example,the zero-forcing(ZF)decoding,minimum mean-square error(MMSE)decoding and maximal ratio combining(MRC)decoding,can achieve high efficiency and performance,computational complexity is still high due to the requirements of the inverse of the channel matrices,as the number of antennas increases.For this reason,to find a decoding method with more low complexity is a meaningful work.In this thesis,two decoding methods with low decoding complexity are proposed based on the characteristics of a massive MIMO system,one is called as fast decoding method and the other is named as simple operator decoding method.The fast decoding method is based on the asymptotic orthogonality of the channels in the massive MIMO system.In this thesis,two cases are considered.One is under assumption that the decoding side is fully aware of the channel state information.In this case,the inverse matrix of the channel matrix is not necessary in our proposed decoding method,by using the asymptotic orthogonality of the massive MIMO system.In fact,in the method,problem of joint decoding is transformed into KL sub-problems with very low decoding complexity,where K is the number of users in each cell,and L is the number of cells in the system.Thus,the computational complexity is greatly reduced.The other is the case where the channel state information is assumed to be unknown.For this case,we also use the asymptotic orthogonality of the channel,and let the minimum value problem of the joint decoding method to be transformed into K sub-problems.Also,calculation of the inverse matrix of the channel matrix is not needed.Hence,decoding complexity is also greatly reduced.The simple operator decoding is based on the following characteristics: as the number of antennas in the base station increases,irrelevant interference and additive noise among the users are asymptotic orthogonality.Based on this feature,this thesis presents a low-complexity decoding method based on pilot transmission.This method also has low decoding complexity by combining the received base-band observation matrix and the received data signal matrix together.The method avoids the estimation of the channel matrix and the calculation of the inverse matrix of the channel matrix.Simulation results show that the proposed methods greatly reduce the complexity of the estimation and decoding.Moreover,the resulting in performance loss is in an acceptable range.