Research On Channel Estimation Of MIMO Communication System

The rapid growth of mobile data services has brought great challenges to wireless access.In the future,mobile communications need to make more efficient use of broadband resources to greatly improve spectrum efficiency.Multiple-Input Multiple-Output(MIMO)and massive MIMO technologies,through the multi-antenna characteristics of both transmitters and receivers,can greatly improve the system channel capacity without increasing spectrum resources.They are regarded as the core technologies of mobile communications.Channel estimation technology based on MIMO communication system is an important subject in this field.The performance of MIMO communication system is directly affected by the advantages and disadvantages of channel estimation.To improve the communication quality of MIMO system,this paper focuses on the design and optimization of training sequence,channel estimation based on training sequence,channel estimation based on compressed sensing and channel estimation based on angle domain.The main contributions of this paper are as follows:1.A training sequence optimization algorithm based on maximizing mutual information is proposed.Based on orthogonal Gold sequence,Zadoff-Chu sequence and complete complementary sequence,the training sequence is optimized under the condition of Gaussian white noise and clutter.The basic training sequence is operated by matrix operation by spectral decomposition to satisfy the optimized sequence.Orthogonality and channel capacity of MIMO communication system based on the optimized sequence are simulated.The results show that the optimized orthogonal Gold sequence can achieve better channel capacity under both white Gaussian noise and clutter conditions,which verifies the feasibility of the optimization algorithm.2.A hybrid MIMO channel estimation algorithm based on multivariate training sequences is proposed,which combines the characteristics of sequential training and superposition training.The channel estimation accuracy and system performance were improved using the excellent auto-correlation function and the low peak-to-average power ratio of multivariate training sequence in timing sequence insertion.Specifically,in order to further enhance the utilization of high frequency spectrum,the training sequence in the superposition method does not occupy a single time slot.The channel state was estimated by enhancing the training sequence signals through the periodic superposition of multivariate training sequence.On this basis,the power and length of the training sequence were adjusted flexibly through the integration between the channel estimation model based on timing sequence insertion and that based on superposition,thereby increasing the spectrum utilization of the MIMO communication system.3.A method of constructing deterministic measurement matrix based on correlation criterion is proposed,and the measurement matrix is applied to compressed sensing massive MIMO model to study channel estimation algorithm.By studying the conventional massive MIMO channel model,the corresponding compressed sensing channel model is obtained.The model echoes the function of pilot sequence in massive MIMO channel and the function of measurement matrix in compressed sensing signal model.Different measurement matrices are used to study the massive MIMO channel estimation algorithm and accurate signal reconstruction is realized.The channel estimation models based on compressed sensing,LS and different measurement matrices are compared and analyzed.The results show that the channel estimation algorithm based on the complete complementary sequence as the measurement matrix reduces the complexity of the system while improving the performance of channel estimation compared with other measurement matrices.4.A least squares ESPRIT algorithm based on mixed precoding is proposed.This algorithm extracts the channel characteristic parameters of massive MIMO system by estimating the angle information of transmitter and receiver.At the same time,it uses the hybrid analog-digital precoding method to compensate for the high-power load of RF link in all-digital receiving array system.Based on the normalized mean square error evaluation criterion,the theoretical formula of channel estimation for least squares ESPRIT algorithm with hybrid analog-digital precoding was deduced.The simulation results show that the proposed algorithm can reduce the computational complexity while guaranteeing the estimation accuracy.