Robust Cooperative Guidance of Leader–Follower AAVs Against Maneuvering Target With False-Data Injection

To address the cooperative guidance problem for autonomous aerial vehicle (AAV) formation with false-data injection and only the leader equipped with a seeker, a hybrid algorithm is developed to estimate the line-of-sight (LOS) angles between AAVs and the target by integrating Tiny Long Short-Term Memory (TLSTM) and Kalman Filter (KF). A cooperative guidance method with the leader-first follower (LFF) configuration is proposed to achieve simultaneous attacks on a maneuvering target within a prescribed time by regulating the contraction rate of the formation. First, the Expectation-Maximization (EM) algorithm is employed to optimize the parameters of the KF, while the TLSTM model is integrated for updating the observation for the optimized KF. The KF extracts local features of the angular data to smooth the angle predictions from the TLSTM. Subsequently, a prescribed-time guidance law is developed for the leader in the LOS direction. To ensure rapid convergence of follower AAVs, a prescribed performance control method is adopted to adjust the contraction rate of the formation. Finally, rigorous theoretical demonstration and numerical simulations validate the effectiveness of the proposed approach.