Intelligent model correction and trajectory planning for air-breathing hypersonic vehicle considering inlet unstart

This study develops a protection mechanism against inlet unstart for air-breathing hypersonic vehicles (AHVs) by predicting potential unstart scenarios using a mechanism model. A deep neural network (DNN)-based trajectory planner is employed to avoid unstart-triggering flight paths. AHV mechanism model is described, and error sources are analyzed. A reliable sensor feedback scheme is designed to correct model parameters using neural networks. The trajectory optimization problem is formulated as a highly nonlinear optimal control problem, with state-action vectors extracted from optimal trajectories generated from random initial states. A DNN is then trained to learn the relationship between flight states and optimal actions, enabling optimal action prediction. The key contribution of this study lies in integrating neural networks with mechanism model correction and trajectory optimization to prevent inlet unstart. The algorithm's effectiveness is validated through numerical simulations.