Robust control through signal constraints with application to predictive control

Robust control through signal constraints with application to predictive control

de Vries, R.A.J.

Verhaegen, M. (promotor)

Mechanical Maritime and Materials Engineering

2006-12-18

This thesis presents a novel approach to robust controller design. It describes how linear constraints can be derived on the output of an arbitrary controller, that guarantee robust bounded-input/bounded-output stability and asymptotic perfect tracking of the reference when the external signals become constant. The results are applicable to a large class of multivariable, non-linear, time-varying, perturbed processes. The constraints require minimal information about the possible perturbations; only an upper bound on their l1-norm. Still, almost all available knowledge about the perturbations can be incorporated in a transparent and optimal way. This is possible by the introduced concept of disturbance detection: all knowledge about the perturbations is used to distinguish the effect of disturbances from the effect of perturbations. The derived constraints are made unconservative by taking two things into account: the detected effect of disturbances and the difference between the measured effect of the perturbations and their maximally allowed worst case effect given the desired robust performance. The constraints incorporate anti-windup, are easy to implement and can be transparently tuned between good nominal and robust performance. For open-loop stable systems the constraints are guaranteed feasible in the presence of time-varying hard level and rate constraints on the controller output. The constraints are applied to a predictive controller, in which case they are tuned automatically depending on the settings of this controller. Several simulation studies confirmed that the developed robustness constraints and tuning rules performed well.

robust control
predictive control
non-linear control
optimal control
offset-free tracking
signal constraints
contraction constraints
linear constraints
robust stability
disturbance detection

http://resolver.tudelft.nl/uuid:01d2a070-d657-4aa3-b008-965121a34bdd

90-9021429-1

Institutional Repository

doctoral thesis

(c) 2006 R.A.J. de Vries