<p>A DualBiNet model for human behavior recognition using Channel State Information (CSI) signals is presented in this study. The model aims to provide an effective solution for accurately recognizing diverse human behaviors in specific environments through deep learning techniques. First, we review the relevant knowledge and recent research trends in the field of human behavior recognition utilizing CSI signals, highlighting their potential applications. Second, we detail the preprocessing techniques employed in our study to enhance data quality and model performance. Subsequently, we explore the architecture of the DualBiNet model and its unique advantages, including the effective combination of spatial feature extraction and temporal sequence modeling. Finally, we evaluate the model’s performance on two widely recognized public datasets, achieving recognition accuracies of 99.03% and 98.47%, significantly surpassing other methods. These results underscore the effectiveness of the DualBiNet model for human behavior recognition using CSI signals, providing valuable insights for future research in this domain.</p>