Dynamic behavioral modeling of 3G power amplifiers using real-valued time-delay neural networks

In this paper, we propose a novel real-valued time-delay neural network (RVTDNN) suitable for dynamic modeling of the baseband nonlinear behaviors of third-generation (3G) base-station power amplifiers (PA). Parameters (weights and biases) of the proposed model are identified using the back-propagation algorithm, which is applied to the input and output waveforms of the PA recorded under real operation conditions. Time- and frequency-domain simulation of a 90-W LDMOS PA output using this novel neural-network model exhibit a good agreement between the RVTDNN behavioral model's predicted results and measured ones along with a good generality. Moreover, dynamic AM/AM and AM/PM characteristics obtained using the proposed model demonstrated that the RVTDNN can track and account for the memory effects of the PAs well. These characteristics also point out that the small-signal response of the LDMOS PA is more affected by the memory effects than the PAs large-signal response when it is driven by 3G signals. This RVTDNN model requires a significantly reduced complexity and shorter processing time in the analysis and training procedures, when driven with complex modulated and highly varying envelope signals such as 3G signals, than previously published neural-network-based PA models.