RNN Online Learning-Based Fixed-Time Prescribed Performance Formation Control for Networked Heterogeneous HFV Under Event-Triggered Communication

This article investigates online learning-based fixed-time prescribed performance collaborative formation control for swarm heterogeneous hypersonic flight vehicles (HFVs) with multiple constraints (i.e., unknown aerodynamic perturbations, system unmodeled and parameter uncertainties, as well as unknown communication link faults) under the event-triggered communication mechanism (ETCM). A fully distributed fixed-time event-triggered estimator (FDF-TE-TE) is first presented to estimate the state of a virtual leader, which simultaneously considers many factors, including event-triggered communication, communication link faults, fixed-time convergence, without the global knowledge of communication topology, and unknown inputs of the virtual leader. To tackle the problem of heterogeneous HFV swarms' unknown terms caused by disturbances and uncertainties, recurrent neural networks (RNNs) are constructed for each HFV to online learn and approximate them. Next, based on the designed FDF-TE-TE and compensation terms of the RNNs, a fixed-time prescribed performance cooperative formation control protocol (F-TPPCFCP) is proposed. It not only makes the fo-rmation errors converge to the preset residual range within any time user-expected but also eliminates the need to calculate their initial value, thus overcoming the drawbacks of existing prescribed performance controls (PPCs). Finally, two numerical simulations are conducted to illustrate the effectiveness and superiority of the proposed algorithm.