Research On Digital Predistortion Algorithm Based On Memory Polynomial

As an important part of the composition of the wireless communication transmitter, the power amplifier status is beyond doubt. In recent years, the rapid development of wireless communications, the frequency of the spectrum of resources to make use of a higher frequency modulation signal to become mainstream. These modulation signals are very sensitive to the nonlinear power amplifier. To ensure the quality of the signal, the linearity of the power amplifier is put forward higher requirements, so the linearization technology is booming. In these technologies, baseband digital pre distortion due to the low relative cost, the optimization effect is good, so instead of feed forward, become now the mainstream of the commercial base stations in the linearization technology.The content of this paper is focused on the research of the pre distortion algorithm and the verification of the platform. Different from the now widely used indirect learning architecture, learning structure used in this paper is divided into two steps: firstly, the model of the power amplifier is calculated, through the parameter extraction algorithm specific parameters of the power amplifier model; then through analysis on the structure of the model, so as to determine the pre distorter the overall structure, the parameters are computed. This paper uses memory polynomial to model the power amplifier, and the predistorter is memory polynomial inverse model. In the process of inverse model, according to the actual situation of the inverse polynomial LS algorithm based on the algorithm can get better fitting results according to the actual situation.This paper is put forward by using technology to build digital predistortion instrument verification platform, the program includes the following instruments, vector signal analyzer(VSA) and vector signal generator(VSG). The use of the platform for a job in the class AB power amplifier is optimized for 95 MHz. The proposed algorithm in the same output power under the condition of the bandwidth of 1MHz signal of linear OFDM improved more than 15 dBc. The test also proves that the fitting algorithm is proposed in this paper is compared with 3dB improved the traditional LS algorithm improvement.