DOI码：10.1109/TMTT.2025.3571780

所属单位：图书馆VIP信息学院

教研室：电子工程与信息科学系

发表刊物：IEEE Transactions on Microwave Theory and Techniques (Early Access)

项目来源：NNSF 62371436

关键字：Digital predistortion (DPD), learnable activation function, neural networks (NNs), power amplifiers (PAs).

摘要：In some application scenarios, radio frequency (RF) devices face strict power consumption limits, necessitating digital predistortion (DPD) models with lower complexity. To address the needs of low-complexity scenarios, a novel DPD model called learnable edge-located activation neural network (LEANN) is developed in this article. Unlike traditional neural network (NN) models that use a uniform activation function at the nodes, the core idea of the proposed LEANN model is to enhance the flexibility and interpretability of nonlinear modeling by employing learnable univariate functions as activation functions on the edges of the network. Furthermore, given the varying nonlinear characteristics of di erent signal components, a logarithmic regularization pruning method suitable for the LEANN model is also proposed. This method promotes a greater sparsity in the model by reducing the similarity between activation functions. Experimental results demonstrate that the proposed LEANN model achieves a lower complexity and higher performance compared to several classic linear parameter models and NN models in linearizing power amplifiers (PAs). Furthermore, the pruned LEANN(PLEANN)modelfurther reduces the complexity without significantly decreasing the performance.

第一作者：Junsen Wang (王俊森)

合写作者：Renlong Han,Qianqian Zhang,Chengye Jiang,Hao Chang

论文类型：期刊论文

通讯作者：Kang Zhou,Falin Liu

论文编号：10.1109/TMTT.2025.3571780

学科门类：工学

文献类型：J

卷号：Online

期号：Online

页面范围：1-14

是否译文：否

发表时间：2025-06-05