Research On The Low Complexity Sparse Channel Estimation Algorithms Based On Hierarchical Model For SC-FDE System

With the growth in demand for all kinds of Internet multimedia and the rapid development of network technology, wireless communication receives increasing attention from scholars. As one of the key technology of the fourth generation(4G) wireless communication system, single carrier frequency domain equalization(SC-FDE) on the one side, overcome the disadvantages of OFDM system, on the other side, compared with the OFDM system, it must keep with the same complexity of the OFDM system in order to obtain the same performance with it. The frequency domain equalization and the detection of the receiver need channel state information, so the channel estimation is an essential part of the SC-FDE system.In the traditional pilot aided channel estimation method, the frequency interval between pilot subcarriers is generally required less than the coherent bandwidth, only in this way can we use the interpolation method to get the impulse response of the data more accurately. While,this lows the spectrum efficiency of the system. However, in actual communication system, we often encounter the sparse channel(such as: broadband wireless communication channel, underwater acoustic communication channel, etc.), the impulse response of the sparse channel only contains a small amount of larger amplitude values, and also the interval between these values is considerably larger. This makes the channel seems to be sparse. In the traditional channel estimation method, people always overlook the sparse character of the channel. Therefore, this thesis will focus on sparse channel estimation methods for the SC-FDE system.The main work of this thesis is following described:First we briefly introduce the development of wireless communication and the technology of SC-FDE system.Then we analyze the characteristics of the wireless communication channel, and the current research status of the channel estimation for SC-FDE system. Finally, we describe factor graphs and message passing algorithms, and give the calculation formulas for different kinds of message update rules.We then make researches on the factor graph models of the 3-L Bayesian hierarchical model.A kind of partitioned MF algorithm is proposed based on the 3-L vector form graph model. The simulation analysis its performance in terms of convergence rate and mean square error compared to the original unblocked MF algorithm and shows its rationality in theory.In the improved scalar 3-L hierarchical model, we use GAMP algorithm instead of BP algorithm and proposed GAMP-MF algorithm based on the improved scalar 3-L hierarchical model. And finally, the lower complexity algorithm GAMP-MF-ELI is raised based on the GAMP-MF algorithm.Simulation is made to show its performance compared to the unblocked MF algorithm.Finally, the proposed partitioned MF algorithm based on the vector form of the 3-L hierarchical model and GAMP-MF-ELI algorithm based on the modified scalar form of the 3-L hierarchical model are applied to the channel estimation for SC-FDE system. Monte-Carlo simulation is used to evaluate the performance and complexity of the raised algorithm when compared to the classic greedy algorithm OMP and the state of art algorithm unblocked MF.