DL-Based Reduced-Order Modeling and Fuzzy Output Tracking Control Design With Vibration Suppression for a Flexible...

Area of Science:

• Aerospace Engineering
• Control Systems Theory
• Computational Dynamics

Background:

• Flexible variable wingspan aircraft (FVWA) exhibit complex rigid-flexible coupled dynamics.
• Accurate modeling is crucial for effective control and vibration suppression.
• Existing methods struggle with the high dimensionality and complexity of FVWA dynamics.

Purpose of the Study:

• To develop a reduced-order modeling framework for FVWA.
• To design a finite-dimensional guaranteed cost output tracking (GCOT) control for vibration suppression and flight command tracking.
• To provide an upper bound for the quadratic flight performance index.

Main Methods:

• Integration of deep learning (DL) for dimensionality reduction and sparse identification of nonlinear dynamics (SINDy) for governing dynamics identification.
• Development of a Takagi-Sugeno (T-S) fuzzy control algorithm combined with integral control.
• Utilizing ordinary differential equations (ODEs) and the beam equation to represent system dynamics.

Main Results:

• A successful reduced-order model for FVWA was established.
• The proposed fuzzy control algorithm effectively achieved flight command tracking and fuselage elastic vibration suppression.
• Numerical simulations validated the theoretical method's effectiveness and superiority, providing an upper bound for the performance index.

Conclusions:

• The combined DL and SINDy approach provides an effective solution for reduced-order modeling of complex FVWA dynamics.
• The finite-dimensional fuzzy control strategy ensures robust performance in flight command tracking and vibration mitigation.
• This research offers a significant advancement in the control design for flexible aircraft with variable wingspans.