Finite-Time Deterministic Learning Command Filtered Control for Hypersonic Flight Vehicle

Abstract

In this article, the finite-time (FT) deterministic learning control for the hypersonic flight vehicle (HFV) dynamics with model uncertainty is investigated. The design is divided into an offline training phase and an online control phase. First, in the offline training, the radial-basis-function neural networks (RBF NNs) are set along the periodic signals to guarantee the partial PE condition. Meanwhile, the offline FT composite learning laws are constructed driven by the system tracking and learning performance index. Embedding the FT composite learning in the FT command filtered control framework, the FT convergences of the system tracking and learning are guaranteed simultaneously. Moreover, the near-optimal learning knowledge is stored. In the next online process, the stored NNs weights are directly used in the online tracking controller without repeatedly updating the weights. Simulation on HFV dynamics shows that the offline FT learning control can achieve better learning and tracking performance, while recalling the stored knowledge online not only guarantees the control performance but also reduces the computational load.