Research on Health Management and Fault Prediction for Human/Unmanned Collaborative Cluster Flight Control System Based on GAN-LSTM

With the rapid development of unmanned aerial vehicle (UAV) technology, manned-unmanned cooperative cluster plays an important role in modern warfare. In order to achieve the health management and fault prediction function of the cluster through the collaborative flight control systems, this study proposes an improved anomaly prediction model GAN-LSTM that integrates long short-term memory network (LSTM) and generative adversarial network (GAN). The model can capture the dynamic characteristics and long-term dependencies of time series data through the LSTM network, while utilizing the generative and discriminative mechanisms of GAN to learn the complex distribution of data and enhance the recognition capability of abnormal fault patterns. This study conducts a series of comparative experiments between the GAN-LSTM model and traditional models like OC-SVM, GRU, and LSTM, focusing on critical performance indicators such as detection rate, accuracy, precision, F1 score, and false positive rate. The results indicate that the GAN-LSTM model surpasses traditional models in detection rate, accuracy, and F1 score, while maintaining high precision and a low false positive rate, achieving an anomaly prediction accuracy rate above 95%. This research offers substantial technical support for the aerial status monitoring and safety assessment of UAV clusters, significantly contributing to the advancement of intelligent and automated cooperative operations for both manned and unmanned clusters.