As the key components of the transmitter, the linearity performance of Shortwave Power Amplifier has a significant influence on the signal quality with restrict linearity requirements of the shortwave wireless communication(1.6MHz~30MHz). At present, the Power Amplifier linearization technique has gradually become a hot research issue. The impressed characters of digital pre-distortion(DPD) technique, including the flexible structure and wide bandwidth from the various types of linearization technique, has made it become the most promising linearization technique.This thesis studies the static and dynamic characteristics of the power amplifier and various kinds of digital pre-distortion structures. Applying the features of the shortwave communication as the background, the writer chooses the direct learning structure to further improve the performance of pre-distortion linearization. And meanwhile, this thesis explores the influence of pre-distortion with time-varying characteristics optimized by interrelated algorithms on the power amplifier.As for the realization of digital pre-distortion, this thesis selects the memory polynomial as the basic model. Compared with the hardware of conventional predistortion which implemented only by FPGA, this proposed pre-distortion combines FPGA and DSP which realize the structure and parameter respectively. This method makes the digital pre-distortion more scalability, applicability and improves efficiency when designing DPD. At the same time, the key technologies when implementing the hardware, including finite word-length effect, look-up table(LUT) technology and cordic algorithm for mold, have been analyzed in detail.In the procedure of validating and testing the digital pre-distortion, the impacts on signal acquisition accuracy from channel carrier frequency offset, mirroring frequency and carrier leakage have been analyzed in the process of signal acquisition. Finally, the digital pre-distortion algorithms are simulated in PC and tested by the vector signal generator(VSG) and spectrum analyzer. The tested results reveal that the proposed digital pre-distortion can make the IM3 of output signal improved more than 20 dB and the IM3 can even reached 28 dB at some frequency points with optimization in the shortwave communications system band. In the whole band of the shortwave communications system, the IM3 of the output signal can be less than-50 dB in linear degree. Results show that this digital pre-distortion can improve linearization of the shortwave communication system effectively. |