Research On Channel Estimation And Channel Equalization Of Filter Bank Multicarrier System

As one of the promising candidates for 5G's new waveform technology,Filter Bank Multicarrier has been paid much attention.Compared with Orthogonal Frequency Division Multiplexing,it uses a new type of filter with good time and frequency localization,which can resist interference well without cyclic prefix and meanwhile obtain a higher spectral efficiency and lower out-of band radiation.However,all of these advantages are at the expense of relaxing the orthogonal conditions of subcarriers,and this non-orthogonality brings for FBMC system the inherent interference that can't be eliminated,so that the channel estimation and channel equalization methods in FBMC system are more complicated than methods in OFDM system.Based on the above reasons,this dissertation researches the channel estimation and channel equalization techniques of FBMC system.Firstly,the thesis studies channel estimation techniques of the FBMC system.In the aspect of channel estimation in the frequency domain,the thesis mainly focuses on the zero-setting method and interference cancellation method in the interference elimination method,Interference Approximation Method in the interference using method and Pairs of Pilots in the interference avoidance method.Among the three methods,IAM has the best estimation performance while the POP method gains the worst,and due to the approximate equality of the channel,these three estimation methods all have error floor.In view of channel estimation in the time domain,the Linear Minimum Mean Square Error method and Weighted Least Square method are researched primarily.Simulation reveals that channel estimation methods in the time domain have better result than methods in the frequency domain because the former have no additional limit to the channel.Secondly,the dissertation studies channel equalization techniques of the FBMC system,including Zero Forcing method,Minimum Mean Square Error method and Equalization with Interference Cancellation method.Both the ZF and MMSE methods have performance platforms,while EIC method has a better performance than traditional equalization algorithms by eliminating the inherent interference of the FBMC system.This thesis proposes an improved algorithm called MMSE-EIC,which can further improve the performance of the system.The simulation results demonstrate that the developed equalizer has a better effect,and the performance can be further optimized by different iterations.