Extending an approximate modelling framework for robust control system design to include periodic input data

Describes a procedure under development for going from plant test data to plant descriptions compatible with the existing robust control design framework comprising of an approximate model plus uncertainty bound combination. Although conceived for application to the periodic input data case, the procedure can in principle be employed with any form of plant response. The new scheme uses fairly standard parametric identification method but also draws fundamental features from a new systematic theory of system identification error analysis developed by Ljung (1987). The authors describe how this theory is used to construct a mechanism for computing uncertainty bounds for identified transfer function models