Maximilian Mühlegg

Learn More
In Model Reference Adaptive Control (MRAC) the modeling uncertainty is often assumed to be parameterized with time-invariant unknown ideal parameters. The convergence of parameters of the adaptive element to these ideal parameters is beneficial, as it guarantees exponential stability, and makes an online learned model of the system available. Most MRAC(More)
A concurrent learning adaptive-optimal control architecture for aerospace systems with fast dynamics is presented. Exponential convergence properties of concurrent learning adaptive controllers are leveraged to guarantee a verifiable learning rate while guaranteeing stability in presence of significant modeling uncertainty. The architecture switches to(More)
We study direct model reference adaptive control of linear systems with noisy measurements. The focus is on achieving robustness of a concurrent learning adaptive controller against noisy measurement. Concurrent learning adaptive control uses specifically selected and online recorded data concurrently with instantaneous data and is capable of guaranteeing(More)
A concurrent learning adaptive-optimal control architecture is presented that combines learning-focused direct adaptive controllers with model predictive control for guaranteeing safety during adaptation for nonlinear systems. Exponential parameter convergence properties of concurrent learning adaptive controllers are leveraged to learn a feedback(More)
A method based on Bayesian linear regression for output monitoring of an adaptive controller is presented. As a basis, a feedback linearized system is augmented by a Model Reference Adaptive Controller. The application of Bayesian linear regression with online recorded data allows the prediction of the adaptive control output and the detection of(More)
Optimal control of autonomous aircraft with modeling uncertainties is a challenging problem, especially considering that onboard computational resources may be limited. A concurrent learning direct model reference adaptive control architecture with reference command optimization is presented. Exponential parameter convergence properties of concurrent(More)
We present a new algorithm for GP regression over data with non-Gaussian likelihood that does not require costly MCMC sampling, or variational Bayes optimization. In our method, which we term Meta-GP, we model the likelihood by another Gaussian Process point-wise in time. This approach allows for the calculation of the posterior predictive mean and variance(More)
Model Reference Adaptive Control facilitates nonlinear systems to adapt to modeling errors, environmental changes or structural damage. Most adaptive control frameworks employ Lyapunov analysis in order to establish stability of the closed loop system. However, the derived signal bounds are inherently conservative. Furthermore, these bounds are seldom(More)
  • 1