Stability Margins with Real Parameter Uncertainy: Test Data Implications

A stability robustness problem is posed in which a plant description contains both parametric and unstructured uncertainty, and test data is available for reduction of the parameter uncertainty through identification. An expression is obtained for the minimal stability margin consistent with the test data and unstructured uncertainty assumptions. Computational procedures for obtaining bounds on the stability margin are indicated. The computations require the solution of a convex optimization problem.