Certification of Adaptive Control for Manned and Unmanned Aircraft - Carnegie Mellon Silicon Valley-Silicon Valley Campus - Carnegie Mellon University

Certification of Adaptive Control for Manned and Unmanned Aircraft

nonlinear simulation
Figure 1: Nonlinear simulation environment
(Reflection) to test UAV control architectures.

A significant challenge in flight certification is the linkage between existing certification metrics for linear control with the nonlinear adaptive control methods developed over the last thirty years. These advanced control methods have demonstrated the ability to enhance the safety of aircraft in the event of damage in flight. To overcome this gap, a new foundation for adaptive control design and stability analysis must be pursued as the traditional nonlinear framework is not suitable.

The current research thrust specifically entails:

  • analysis of time-delays in the adaptive loop using the Razumikhin method and the Lyapunov- Krasovskii functional approach;
  • optimization methods for time delay margin extraction;
  • an extension of phase margin for LTV systems using Joint-Time Frequency Methods;
  • an extension of phase margin using Joint-State Frequency Methods.

We are currently applying the above techniques to inner-loop, outer-loop and trajectory following control of aircraft.

Contact: Abe Ishihara

outerloop diagram
Figure 2: Inner and outer-loop control framework.