A Hybrid Resnet-Bilstm Network Approach for Airborne Fire-Control Radar Operational Mode Recognition

To address the limitations of traditional radar signal processing methods in achieving sufficient recognition accuracy for operational modes under complex electromagnetic environments, this paper innovatively proposes a deep learning approach based on a ResNet-BiLSTM hybrid network. This method introduces typical algorithms from the visual image recognition domain into radar signal analysis. Its core innovation lies in thoroughly analyzing the temporal-spatial coupling characteristics exhibited by signals from different radar operational modes. By constructing a dual-stream parallel fusion network, the approach effectively integrates temporal sequence information and spatial features of radar signals, enabling precise identification of airborne fire-control radar operational modes in challenging environments. Simulation experiments comprehensively validate the effectiveness of the proposed method. The ResNet-BiLSTM hybrid network demonstrates exceptional recognition accuracy, achieving a significant improvement over conventional single-network architectures such as VGG16, ResNet18, and BiLSTM, with stable performance consistently exceeding 95 %. The results highlight the method's capability to address the critical challenge of operational mode recognition in modern electronic warfare scenarios.