Research And Implementation Of Channel Estimation In NC-OFDM Systems

The update of communication technology is getting faster and faster,people have increasingly high requirements for reliability and efficiency of information transmission.Non-Continuous Orthogonal Frequency Division Multiplexing(NC-OFDM),as a physical layer data transmission solution based on OFDM technology,can avoid the interference and reuse the spectrum.Not only guarantee the reliability of information,but also effectively improve the information transfer rate,so it has been widely concerned by scholars.Channel estimation algorithm is one of the key technologies of NC-OFDM system.In this paper,the problem of DFT algorithm for signal-to-noise separation is not clear and the low estimation accuracy of block LMMSE algorithm will be stduied,we propose new solutions and verify the feasibility of the new algorithm.The main research contents are as follows:First,this paper discusses the application background of NC-OFDM system in software defined radio,clarifies the importance of channel estimation technology to NC-OFDM system,and introduce the development trend of channel estimation technology.After that,the theoretical models of wireless channel and OFDM system are introduced in detail,and the suitable pilot pattern is selected through the simulation analysis.Second,the performance and complexity of several classical channel estimation algorithms are simulated and analyzed.Aiming at the problem that the signal-to-noise separation in the DFT algorithm can't adapt to the channel variation,a K-DFT algorithm based on K-means is proposed to separate signals and noise by clustering.The simulation shows that the algorithm has better estimation performance than the DFT algorithm,and it is conducive to project implementation.Then,aiming at the problem of neglecting part of the autocorrelation information in partial LMMSE algorithm,this paper proposes an improved LMMSE algorithm called T-LMMSE.The new algorithm decomposes the autocorrelation matrix of the channel and preserves the channel autocorrelation information to the maximum extent,then,constructing vector of cyclic autocorrelation matrix for cyclic preconditioning matrix.The results of MATLAB simulation verify that the new algorithm reduces the complexity of the algorithm while ensuring the accuracy of the algorithm estimation.Finally,this article has carried on the project to the traditional DFT algorithm and the improved DFT algorithm on XILINX FPGA chip.And built a complete communication link based on NC-OFDM system.By comparing the transmission results of LS algorithm with the improved DFT algorithm and the constellation diagrams before and after the channel estimation module,it is verified that the new algorithm still has good channel estimation performance under the premise of easy implementation of the project.