The Channel Estimation Algorithm And Codebook Design In MMWave MIMO System

With the development of communication technology and internet,the demand in data rate of software and application increases rapidly.For now,the achievement of the virtual reality and large-scale interactive network technology is not just limited by the hardware but also by the wireless data rate.Therefore,it is a hot spot of the present research to provide a high speed and high quality communication network system.The fifth generation mobile communication technology will be excepted to enter the commercial operation phase in 2020.The main technology in 5G,such as millimeter wave band and massive MIMO,are not only technological innovations,but also brings many challenges.The most important different of massive MIMO system and the traditional system is the number of antennas used is dramatically increasing.These large-scale antennas,on the one hand,bring the improvement of the system performance,on the other hand,also makes the dimension of the channel matrix increase dramatically,which means it is more difficult to estimate the channel in a limited time.The communication frequency of the fifth generation communication system is mainly concentrated in the millimeter wave band and the millimeter wave have sparse features in the angle domain.Using this sparse property,scholars propose a channel estimation method based on compressed sensing in millimeter wave band.Which has broken through the limitation of Nyquist sampling rate,and can recover the original data with the probability of large probability when the sampling rate is much lower than double the maximum frequency.This provides theoretical basis and a solution for channel estimation in frequency division system.In this paper,some problems in the channel estimation based on compressed sensing are discussed and solved.The main achievements of this paper are as follows.In the third chapter,we have summarized the theoretical basis of compressed sensing in channel estimation,such as adaptive channel estimation,matching pursuit in hybrid beamforming design and orthogonal matching pursuit algorithm.Based on these theories,we further propose the channel estimation scheme under the condition of channel path unknown in the fourth chapter,and channel estimation scheme based codebook updating in the fifth chapter.These two schemes solve the following problems respectively:First,in the traditional channel estimation algorithms,the methods usually need the system to provide the number of paths in the channel,which means the number of reachable paths from the base station to the mobile station in the channel.Due to the high frequency characteristics of millimeter wave,the general reachable paths can be simply considered as two parts:the direct path and the single bounce path.The number of these reachable paths will be severely affected by weather,mobile obstacles in the cell and so on.In the actual system,it is difficult to get this information.Therefore,in this paper,we propose an energy based channel estimation scheme to estimate the channel under the condition of channel paths unknown,which can explore all the available spatial degrees of freedom in the channel as much as possible.Because the channel condition is unknown,the time consumption will be greater than the condition with the number of path known.Therefore,to reduce the overhead,we put forward an estimation method,in which it can reduce the estimation overhead by making use of the redundant information between paths.Secondly,in the multi-path channel estimation,due to the imperfections of the beam waveforms,the paths often influence each other.When estimating the new path,we need to eliminate the estimated paths'contribution to the received energy.In the study,we found that the traditional estimation algorithm generally employs the projection method at the receiver to reduce the contribution of the estimated paths to the received energy.The projection scheme has some deficiencies,one is the mobile station needs to have the antenna configuration information of base station,in order to generate the projection matrix.However,in fact,the antenna arrangement of each base station may vary,this part of the information will be difficult to obtain.Therefore,we need to put forward a method which can reduce the contribution of estimated path to received energy in base station.The second is the dictionary set of beam codewords is over-complete.In the projection method,can only guarantee the residual orthogonal to the latest estimated path,and will introduce in energy in the other estimated direction.Therefore,this paper proposes a channel estimation based on codebook updating to solve the above problems.The main idea is that,after each estimation,the codebook will be updated by using the information of the estimated path in base station,so that the energy is no longer transmitted from the estimated path direction,to reduce the influence of the estimated path.Another advantage of the proposed scheme is that the number of radio frequency links required could be quantified determined instead of setting it through experience.