LSTM-Based Battlefield Electromagnetic Situation Prediction

Abstract

In the modern battlefield, the complex electromagnetic environment has brought considerable challenges to assessing the electromagnetic situation. Traditional electromagnetic situation assessment methods are difficult to process massive high-dimensional data, making it challenging to predict the electromagnetic situation. In recent years, the in-depth development of deep learning has provided a breakthrough in the electromagnetic situation field. However, related research mainly focuses on threat assessment and electromagnetic situation complexity prediction. There are few papers on electromagnetic situation prediction using machine learning. In this paper, we propose an electromagnetic situation prediction method based on long short-term memory (LSTM). We first build an attack-defense model based on deep reinforcement learning to simulate the electromagnetic situation. Then we use LSTM to predict the development of the situation and improve the loss function to reduce the prediction error. Furthermore, we analyze the impact of different situation features on the final win rate and use a small amount of situation information to predict the win rate with high accuracy. The experimental results show that this method can effectively predict the electromagnetic situation, providing excellent decision support for commanders.