Research On Nonlinear Distortion Suppression Technology Of Power Amplifier

As an indispensable part of the wireless communication system,the nonlinear effect of power amplifier can cause distortion of the transmitted signal,which reduces the communication quality of the whole system.In order to compensate for nonlinear distortion,the linearization technique has been extensively studied.With the advantages of intermediate cost,good linearity and portability,the digital pre-distortion technology has gradually become one of the mainstream power amplifier linearization technologies.This thesis first introduces the power amplifier behavior model,indirect learning architecture,direct learning architecture and iterative learning architecture,and then introduces the principle of linearization gain selection,and then studies the parameter identification method in digital pre-distortion technology,while the adaptive distortion algorithm is improved.Finally,this thesis proposes a multi-component linearization method for the receiver.The main innovations are as follows:1)In the indirect learning architecture,in order to solve the contradiction between the convergence speed and steady state performance of LMS algorithm,this thesis proposes a particle swarm optimization(PSO)based LMS algorithm,which uses PSO to update the tap coefficients,which has a better steady-state performance and faster convergence speed.The optimization method is also applicable to the RLS algorithm.On this basis,the PSO-based LMS algorithm is introduced into the structure that is based on Iterative Learning Control(ILC),while the ILC-PSO-LMS algorithm is proposed at the mean time.2)In the direct learning architecture,the second-order Newton algorithm is usually used to complete the parameter identification.For the problem that the algorithm needs a large number of iterations to converge,in this thesis a Newton algorithm combined with particle swarm optimization is proposed.In the iterative process,the position update formula of PSO is used to optimize the tap coefficients and accelerate the approximation of the optimal solution,thus shortening the training time of the pre-distorter weight coefficient.3)On the basis of the originating processing,this thesis proposes a multi-component linearization method of receiver,which first maps the received signal to the high-dimensional space,and models the nonlinear problem of the one-dimensional space as the linear problem of the high-dimensional space.The multivariate linearization method is used to eliminate nonlinear distortion,and finally the decision is demodulated.The performance of the proposed algorithm is verified by a simulation on the AWGN channel.The simulation results show that the proposed multivariate linearization method can achieve the suppression of nonlinear interference.