Honor Paper: Gain-Scheduled Control for Substructure Properties in AMBs

n AMB systems, machine performance can be degraded by the dynamics of the substructure (i.e. the object to which the sensors and actuators are mounted). These dynamics, unlike those of the rotor, actuators, and sensors are not completely known at the time of controller design. This paper introduces a novel approach called a posteriori gain scheduling for addressing this problem and evaluates the suitability of LPV and D-K iteration design methods by applying them to a benchmark problem. An LPV design method modified for time-invariant parameters exhibits performance close to H„ optimal designs, and is modified to account for parameter estimation errors. The resulting robust, gain-scheduled controller exhibits performance close to mixed- n controllers designed using D,G-K iteration.