Low Complexity Digital Predistortion Technology Design And Implementation

In order to improve the spectrum efficiency,OFDM modulation signal is often used in modern communication system.But OFDM modulation signal has higher Peak-to-average power ratio.As an important active device of transmitter,RF power amplifier's linearity determines the overall quality of communication.Baseband digital predistortion technology has become a hotspot for power amplifier linearization with its low cost,high performance and flexible configuration.The main content of this paper is to solve the problem of limited hardware platform resources for predistortion,and propose the design of low complexity digital predistortion.In this paper,low-complexity solutions are given from the aspects of model,integer delay,fractional delay and model parameter extraction.First,the parallel polynomial model is adopted on the model.Compared with the traditional polynomial model,the accuracy of the amplifier is equivalent,but the complexity is lower.Then,the integer and the fractional delay are used in the form of interpolation filter and LMS algorithm,which avoids the complex operation of interpolation,extraction and calculation correlation,thus reducing the computational complexity.Finally,the advantages and disadvantages of the traditional model extraction algorithm are analyzed.Based on this,LC-RLS and LC-RLS-LMS algorithm are proposed,which can reduce the accuracy of the algorithm while maintaining high fitting precision.In addition,this paper builds a digital predistorter based on FPGA,and uses the OFDM signal with bandwidth of 1MHz to test the pre-distortion effect of the model parameters extracted by different algorithms.The test results show that the traditional LMS algorithm with ACPR improves near 15 dB in the case of constant output power.The LC-RLS and LC-RLS-LMS algorithm proposed in this paper have improved the effect of nearly 20 dB,and the pre-distortion effect is equivalent to that of the traditional RLS algorithm,so as to verify the effectiveness of the low complexity predistorter design.