A robust and adaptive MPC formulation for Gaussian process models
Read the original on arXiv Machine Learning →The paper introduces a robust and adaptive model predictive control framework for uncertain nonlinear systems with bounded disturbances and unmodeled nonlinearities, leveraging Gaussian Processes to learn dynamics from noisy measurements. It derives robust predictions for GP models using contraction metrics, integrating them into the MPC formulation to ensure recursive feasibility, robust constraint satisfaction, and convergence to a reference state with high probability. A numerical example involving a planar quadrotor experiencing challenging ground effects demonstrates significant performance gains from the robust prediction method and online learning.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.