Model validation for structured uncertainty models

Model validation concerns the problem of determining if an observed data record is consistent with a given model with prescribed uncertainty bounds. We consider time-domain and frequency-domain validation of linear fractional transformation (LFT) uncertainty models with multiple uncertainty blocks. These structured uncertainty models serve as the basic model for H_∞ and μ-synthesis robust control design. For these uncertainty models, we propose a computationally efficient approximate validation method based on a convex weighted optimization and H_∞ filtering. A simple numerical example is presented