Neural adaptive appointed-time control for flexible air-breathing hypersonic vehicles: an event-triggered case

• Published in: Neural computing & applications Vol. 33; no. 15; pp. 9545 - 9563
• Main Authors: Shi, Yi, Shao, Xingling
• Format: Journal Article
• Language: English
• Published: London Springer London 01.08.2021; Springer Nature B.V
• Subjects: Actuators; Adaptive control; Artificial Intelligence; Communication; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Controllers; Data Mining and Knowledge Discovery; Distance learning; Disturbances; Hypersonic vehicles; Image Processing and Computer Vision; Investigations; Laboratories; Machine learning; Neural networks; Original Article; Parameter uncertainty; Probability and Statistics in Computer Science; Subsystems; Wireless networks; Appointed-time; Flexible air-breathing hypersonic vehicle; Neural adaptive; Event-triggered
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-021-05710-7

This work investigates a neural adaptive appointed-time control for flexible air-breathing hypersonic vehicles subject to modeling nonlinearities, flexible modes, parameter uncertainties and external disturbances. A relative threshold-based neural estimator (RTNE) using minimal learning parameterizations is proposed to online-identify the lumped disturbances with a reduced occupation of communication resource via utilizing intermittent states, while heavy computational burden for online learning is remarkably reduced in the premise of a competitive estimation accuracy. With the estimation results produced by RTNE, a neural adaptive event-triggered control is advanced by incorporating a relative threshold-based sampler into controller-to-actuator channel, such that unnecessary continuous sampling incurring in current time-driven researches can be successfully avoided. Moreover, an appointed-time prescribed performance control is constructed to make the responses of velocity and altitude subsystems evolve within pregiven regions with a user-defined settling time; meanwhile, the strict dependence on the exact knowledge for immeasurable initial system states is removed. The stability of system is proved by virtue of input-to-state stable method, and Zeno behavior is eliminated. Simulations are performed to certify the effectiveness of presented controller.