Quantifying uncertainty in nonlinear systems

Dr G. Vinnicombe (Cambridge University Engineering Dept)

The nu-gap metric provides a means of measuring a "distance" between linear, time-invariant (LTI) systems which captures the importance of any difference between them in the context of feedback system design. In this talk, we give a generalization of the nu-gap which is applicable to nonlinear systems.

Many nonlinear systems are perfectly adequately controlled by LTI controllers. We shall show how an upper bound can be calculated on the nonlinear nu-gap distance between, say, a simplified LTI nominal model and a true nonlinear system. If this distance is small, then the LTI model is perfectly adequate for the purposes of designing an LTI controller for the original nonlinear system.

Back to Control Seminars Page.