基于凸优化的记忆多项式预失真技术

摘要

An efficient adaptive predistorter for high power amplifier(HPA) was designed to improve the convergence rate of adaptive and the robustness.On the base of traditional predistortion techniques,a memory polynomial predistortion based on convex algorithm was proposed in this paper.This algorithm implemented convex optimization using interior point method and avoided the inverse operation of its correlation matrix.As a result,the numerical value was more robust; the operation process was less complex and its speed was faster,in addition,the precision of constringency was better.Finally,two-tone test signals were taken as examples to carry out a simulation.The simulation results show that the algorithm can further provide the adjacent channel leakage ratio(ACLR) more than 5dB,and verify the advantages of the technology.%研究高功率放大器自适应预失真的过程中,为了提高自适应算法的收敛度和稳定度,在传统预失真技术的基础上,提出了一项凸优化算法的多项式预失真技术.利用内点算法来解决凸优化问题,避免了传统RLS算法中对自相关矩阵的求逆运算,提高了数值的稳定性,并且降低了运算的复杂性,提高了运算速度且具有良好的收敛精度.最后,以双音信号为例进行仿真,结果表明,改进算法对邻带交调(ACLR)的抑制至少有5dB的改善,证明改进算法的优越性.