基于混合核的支持向量机数字预失真算法

• 作 者: 宋勇 胡波 李在清
• 机 构: 上海贝尔股份有限公司,复旦大学电子工程系,
• 出 处: 通信技术
• 发表日期: 2013年02期
• 关键词: 数字预失真 记忆效应 核函数 支持向量机 digital predistortion memory effect kernel function SVM (Su

射频功率放大器工作在近饱和点时产生的非线性效应是制约其性能提升的主要因素。提出了一种基于支持向量机的预失真算法。通过引入核函数,可以在高维空间中线性地实现相对于输入空间为非线性的问题,达到更好的预失真效果。采用基于傅立叶核函数与多项式核函数的混合核函数并运用于预失真函数的估计中,提高了算法的学习能力和泛化能力。仿真结果表明,所提算法性能优于传统的基于最小均方的数字预失真算法。

The nonlinear effect that the RF power amplifier operates when it is near the saturation point is the main factor that restricts its performance improvement. A new predistortion algorithm based on support vector machine is proposed. By introducing kernel function, the problem of non-linearity with respect to input space can be achieved linearly in high-dimensional space and better predistortion effect can be achieved. The hybrid kernel function based on Fourier kernel function and polynomial kernel function is applied to the estimation of predistortion function to improve the learning ability and generalization ability of the algorithm. Simulation results show that the performance of the proposed algorithm is superior to the traditional digital pre-distortion algorithm based on least mean square.