Research On Predistortion Technology Of Power Amplifier’s Nonlinear

With the rapid development of mobile communication market, the mobilecommunication technology has been put forward for higher requirements. The last twogenerations communication systems mainly focus on the transmission of traditionalvoice and low rate data service. However, nowadays in the3G and4G mobilecommunication systems, multimedia and high rate data service combined with image,voice, data has become an indispensable communication service issue. Moreover, thenumber of mobile communication users are dramatically increasing. The originalcommunication system has been unable to meet people’s needs, and the developmentbottleneck is mainly the limitation of spectrum resources available. Therefore, for thebetter development of the communication, the utilization of spectrum resources must beimproved. To reach the aim, the linear modulation is more used in communicationsystem at present. But the linear modulation requires the power amplifier has highlinearity, with the inherent characteristics of the semiconductor and the limitation of thelinear region of the power amplifier, the nonlinear distortion can be produced when theamplitude of the input signal is too large. So the baseband predistortion is adopted inthis paper to compensate the distortion caused by the nonlinear characteristic of poweramplifier, so as to improve the linearity of power amplifier.The adaptive principle and algorithm of the baseband predistortion in poweramplifier linearization technique are further studied in this paper, the main contents areas follows:the nonlinear characteristic of power amplifier and its effect oncommunication system performance are introduced, the main linearization methods aredescribed, and the predistortion method is mainly introduced. By analysis andcomparison, the advantages of the baseband predistortion are highlighted.Then, according to analyzing the adaptive principle and algorithm, theHammerstein predistortion system and Wiener predistortion system are established based on adaptive theory. In the Hammerstein predistortion system, a Newton iterativealgorithm, an LS/SVD algorithm, a linear regression algorithm and an improvedvariable step-size LMS algorithm are adopted for pattern recognition, and simulated byusing MATLAB7.1software. The results show that the improved variable step-sizeLMS algorithm proposed in this paper displayed its obviously superiority in out-banddistortion and convergence rate. In the Wiener predistortion system, the linearregression iterative algorithm and the two step regression iterative algorithm areadopted for the system identification, and the disadvantages of Newton iterativealgorithm which performed the sensitivity of estimation on the initial value and the lowaccuracy of linear regression iterative algorithm are overcomed by the two stepregression algorithm, and of which the superiority is embodied on the speed ofconvergence, all these conclusions are verified by simulation.