Neural network parameter identification-based prescribed-time adaptive control for morphing glide aircraft

The morphing glide aircraft (MGA) can adapt to complex environments and mission requirements by altering its aerodynamic configuration through the jettisoned wings, which face the challenges in dynamic and aerodynamic characteristics variation, as well as lumped uncertainty. To deal with the above issues, a neural network (NN) parameter identification-based prescribed-time adaptive control for MGA is proposed. Firstly, a time-scale function is proposed, which avoids the singularity and the unrealistic issue of unbounded growth in control effort. Further, a fractional-power prescribed-time Lyapunov stability theorem is established, which overcomes the limitation of conventional theorems in analyzing robust sliding mode control with non-smooth control terms, providing a theoretical foundation for the design and stability analysis of prescribed-time fractional-power sliding mode controllers. Then, for the issue of aerodynamic parameter uncertainty, an NN parameter identification method is proposed. Based on this, a prescribed-time adaptive sliding mode control of MGA is proposed to guarantee the controller error convergence in the prescribed-time, independent of initial conditions and control parameters. Finally, the proposed algorithm is employed to achieve attitude tracking for the MGA, and the effectiveness is illustrated.