Abstract
The design of robust control for an uncertain dynamical system is first considered. The only information on the uncertainty, which is needed for the design, is its possible bound. The (non-adaptive) robust control design based on this information may sometimes not be feasible or practical. Thus the adaptive version of the non-adaptive robust control is desirable to estimate the bound. Two general routines are constructed such that the corresponding adaptive versions of all the non-adaptive controls can be constructed. The adaptive control is applicable to mismatched uncertainty, measurement noise and the model-following problem. The theory is illustrated via a mechanical manipulator.

This publication has 10 references indexed in Scilit: