BiLSTM-ModernTCN: An Aerial Target Intent Recognition Model

The pervasive integration of advanced technologies in modern military systems has significantly enhanced the informatization of combat platforms, resulting in battlefield situational data characterized by multidimensional complexity and massive volume. By combining bidirectional long short-term memory network(BiLSTM) and modern temporal convolutional network(ModernTCN) to model aerial target temporal intent data, this paper presents the BiLSTM-ModernTCN model. BiLSTM can extract bidirectional temporal features of aerial target intention data, while ModernTCN further captures detailed intent patterns based on conforming to the temporal dependence. The ablation and comparison experiments show that the BiLSTM-ModernTCN model can outperform the existing aerial target intent recognition models.