Studies On Channel Estimation Of MIMO-OFDM Systems

Without introducing any additional bandwidth, MIMO technology can multiply communication system capacity, improve link performance, and increase data throughput of the network. MIMO systems include several mutually independent channels between transmitter and the receiver, and to some degree multipath components of transmission can be used to overcome multipath fading, but the frequency selective fading processing capability is still limited.OFDM system can equivalently divide frequency selective fading channel into several flat fading channels, which can not only create favorable conditions for applications on MIMO in multipath fading channel, but also considerably simplify channel equalization. Based on the assumptions, many domestic and abroad researchers and research organizations have put emphasis on the studies on the combination of MIMO and OFDM technologies in order to achieve higher spectrum utilization efficiency. MIMO-OFDM is a new combination technique of OFDM with MIMO, which uses antenna arrays to realize space diversity and to improve the quality of transmission.This paper carries out the research of channel estimation based on MIMIO-OFDM , and the main work is as following:1. It reviews present conditions of development of wireless and mobile communication systems and introduces their technical features and typical technology of the first generation, the second generation and the third generation (3G) mobile communications are introduced in this paper. Then it discusses the history and current situation of the MIMO systems and OFDM, and provides the perspective of the integration of MIMO and OFDM.2. It discusses the theory of OFDM and channel model of MIMO, and introduces OFDM technology, MIMO wireless communication system model, and MIMO-OFDM channel estimation technology in detail, and which are based on the following chapters.3. It discusses the channel estimation method based on pilots symbol for MIMO-OFDM systems, and studies that transmitting terminal at regular time intervals plug known pilot into the proper position of OFDM symbol, pilot symbol and data symbol together are received by receiving antenna, and by pilot symbol receiving terminal extract channel response at pilot symbol position, then makes interpolation operation by these position information and obtains channel information for the whole period.4. It studies subspace blind channel estimation, for the plugged training sequence or pilot inevitably occupy bandwidth and affect transmitting efficiency of communication systems are affected, and reduce the system effective bandwidth utilization. How to use other observed value including all unkown symbols neglected based on pilot channel estimation algorithm besides pilot symbol to compute channel estimation in order to improve the efficiency of communication system.5. It studies channel estimation based on Bayesian theory, in many problems , required to use observed value with noise to filter and estimate on systems status varying with the time,people often adopt state space method to simulated dynamic system. Bayesian filter theory provide a common frame for dynamic filtering and estimation, and Kalman filter gives the optimal solution for linear and gaussian approximation.But for Non-linear, non-Gaussian problems, it is usually difficult to have analytical form of its solution. So that, all kinds of non-linear filter algorithm are provided. One is extended kalman filter algorithm, which makes a local linearization to a nonlinear system, so Kalman filter can be utilized indirectly to filter and estimate. The other is sequence Monte Carlo algorithm ,that is Particle Filter ,it is the effective methods for non-linear problem s recently appeared .The paper provide a kind of particle filter to realize to channel estimation of MIMO-OFDM system.Main innovations are as follows:1. The channel estimation algorithm based on pilot that adopts equal peak power pilot signal has been proposed. And the dsigned pilot transforms the problem of multiple antennas signal channel estimation into signal antenna situation. Singular value decomposition simplifies computation of inverse matrix in MMSE, and LMMSE algorithm will attain rough estimtion which is as an initial value of the estimation. Then by reducing pilot number and carrying on iterative calculation of the aim function, the proposed algorithm can get more accurate channel estimation.2. In the (semi-)blind channel estimation algorithms in MIMO-OFDM system, the singular value decomposition(SVD) is a useful technique. In this paper, we adopts a adaptive algorithm to estimate the SVD of channel matrix in MIMO-OFDM system, which can fast tracking the signal and noise subspace without eigenvalue decomposition or singular value decomposition. The performance of this method is similar to the source one, but computational complexity is reduction.3. In the practical wireless communication environment, such as the indoor condition or the city environment, there are a mass of non-Gaussian noise. Aiming at the nonlinear/non-Gaussian, the sequential monte carlo particle filter based on Bayesian theorem is proved to be an effective method. In particle filter, the particle degeneration is a key issue to affect the performance of the algorithm. In this paper, we adopt a variable step-size method to regulate the distribution of the resample particle. The simulation result shows that the proposed method can trace the MIMO-OFDM channel variation which affected by the non-Gaussian noise.