Decentralized Robust Adaptive Control for Continuous-Time Linear Plants

Abstract

Results on robustness of continuous-time decentralized adaptive control using a relative deadzone are given for the situation of many adaptive systems interacting through inputs and outputs subject lo static memoryless nonlinear actuators. Stability is preserved by the use of a deadzone function in each adaptation loop. Augmented adaptation error signals taking account of the input deviations from linearity may be used to adapt the adaptive algorithm deadzone to the relative changes of the adaptation error with respect to the estimated unmodeled dynamics contribution.