基于PSO_SVM的射频功率放大器模型研究

Modeling of RF Power Amplifiers Based on PSO_SVM

针对神经网络和支持向量机在射频功率放大器建模领域存在的优缺点,提出一种利用PSO_SVM算法对射频功率放大器进行建模的方法。从理论上分析了支持向量机(SVM)及粒子群优化(PSO)算法的相关原理,并将PSO_SVM算法应用到功放器件建模中。仿真结果表明,基于PSO_SVM的射频功放模型在模型精度、小样本学习和逼近能力方面均优于传统SVM模型和BP神经网络(BPNN)模型。

In view of the advantages and disadvantages of neural networks and support vector machines in application of RF power amplifier modeling, a novel method for modeling RF power amplifier was proposed based on PSO_SVM algorithm. The principle of support vector machine （SVM） and particle swarm optimization （PSO） were theoretically analyzed, and PSO_SVM algorithm was applied to modeling of power amplifiers. Simulation results showed that the RF power amplifier model based on PSO_SVM was superior to traditional SVM model and BP neural network （BPNN） model, in terms of accuracy, small sample learning and approximation capability.