Adaptive fuzzy dominant-pole placement control

The design of an adaptive fuzzy dominant-pole placement controller is addressed. Under a priori known noise specifications, a convex region in the z-domain containing the transfer function parameters can be estimated through application of the set membership recursive least squares identification algorithm. The degree of uncertainty for the estimated parameters is reflected through the area of this region. Fuzzy membership functions are assigned to cells partitioning a universe of discourse, which includes the identified region. The knowledge database consists of feedback gains required to place the closed-loop poles at predefined locations, and a rule-based controller infers the control input variable weighting each gain with the integral of the identified region under the membership functions. The aforementioned controller is demonstrated in simulation studies on an example system