基于Att-BiLSTM的雷达信号调制方式识别

识别雷达信号的调制方式有助于分析雷达的工作模式和目的，为及时采取恰当的应对措施提供依据。长短时记忆网络(Long Short-Term Memory, LSTM)深度学习模型在基于特征的调制方式识别领域中有着广泛应用，但LSTM模型的时间性能会随着输入数据规模的增大而下降。针对以上问题，文中提出了一种基于注意力机制的双向长短时记忆网络（Bidirectional Long Short-Term Memory, BiLSTM）的雷达信号调制方式识别算法。该算法通过BiLSTM提取信号原始数据的特征，再使用注意力机制为学习到的特征分配相应权重，最后由分类器根据学习到的特征输出分类结果。使用Python框架构建基于注意力机制的BiLSTM网络模型，以雷达辐射源信号特征仿真数据作为网络的输入和训练基础，实现对辐射源的调制方式的识别。结果表明，该模型在识别雷达信号的调制方式方面具有良好的效果。

Radar Signal Modulation Recognition Based on Att-BiLSTM

Identifying the modulation mode of radar signals helps to analyze the working mode and purpose of radar, and provides a basis for taking appropriate countermeasures in a timely manner. Long Short-Term Memory (LSTM) deep learning models have wide applications in the field of feature-based modulation mode recognition, but the time perfor-mance of LSTM models will decline with the increase of input data scale. Aiming at the above problems, this paper proposes a radar signal modulation mode recognition algorithm based on Bidirectional Long Short-Term Memory (BiLSTM) based on attention mechanism. The algorithm extracts the features of the original data source of the signal through BiLSTM, and then uses the attention mechanism to assign corresponding weights to the learned features, and finally the classifier outputs the classification results according to the learned features. A BiLSTM network model based on attention mechanism is constructed using Python framework, and the radar radiation source signal characteristic simulation data is used as the input and training basis of the network to realize the recognition of the modulation mode of the radiation source. The results show that the model has a good effect in recognizing the modulation mode of the radar signal.