Research On Channel Estimation Method Of FBMC/OQAM System

5G wireless communication systems need to support applications in a large number of scenarios,which needs to allocate available time-frequency resources flexibly.Therefore,the PHYDYAS project team proposes to use Filter Bank Multi Carrier(FBMC)as an alternative to the key technologies of 5G.In recent years,FBMC technology has gradually become a hotspot in 5G research due to its efficient spectrum utilization and excellent out-of-band characteristics.This paper focuses on the Filter Bank Multi Carrier/Offset Quadrature Amplitude Modulation(FBMC/OQAM)system,and some related research are conducted based on the channel estimation method of this system.The main contents include:First,this paper studies the structure of pilot,and a new pilot interference cancellation method is proposed.This method uses an auxiliary pilot sequence and a precoding algorithm to jointly eliminate interference in the neighborhood of the pilot to ensure a more accurate pilot sequence at the receiver.In addition,an improved algorithm for noise is proposed.The algorithm is divided into two steps.Firstly,a threshold denoising algorithm based on Fast Fourier Transform is used to preprocess the estimated values which obtained Least Square algorithm;Then iterative Linear Minimum Mean Square Error algorithm is used to calculate a more accurate channel estimation value.Simulation results show that the performance of system has been significantly improved.Then,in order to save the bandwidth occupied by the pilot and enhance the tracking ability of the channel,a channel estimation method based on particle filtering is proposed.This method is a semi-blind method.Firstly,the pilot is used to obtain the initial particle set,and then the particle filter algorithm is used to predict the channel estimation value and the observation value is used to modify the estimation result.In the subsequent estimation,the participation of the pilot signal is no longer needed,which saves the system bandwidth.Simulations show that the new algorithm can track the time-varying channel well and improve the performance of the system.Finally,this paper proposes a new estimation method which combine Complex Value Neural Network and channel estimation.This method,proves the feasibility of neural networks in channel estimation.The proposed method is divided into two-phases.The training phase sends the signal from the receiving end of the system with the expected output signal to the network for training,and saves the optimal network parameters when the network converges;During the deployment phase,the optimal network parameters are fixed,and the current Complex Value Neural Network is used to replace the channel estimator part of the original system.Simulation results show that compared with traditional channel estimation methods,the Complex Value Neural Network channel estimation method has higher accuracy.