A robust disturbance-based control and its application

Abstract

A simple and robust disturbance-based control (DBC) method is introduced in this paper which employs a nominal model as the disturbance observer to estimate the overall disturbance, which includes the model-plant mismatch error, unmeasured external disturbances and cross-coupling in multivariable systems. The feedback of the estimated disturbance can eliminate the effect of external disturbances and structured perturbation caused by parameter variation, and decouple a multi-input multi-output (MIMO) system into a number of single-input single output (SISO) nominal subsystems. The root-locus method can be used to analyse the stability of DBC controlled systems and derive the design criterion. Subsequently, a robust controller can be designed based on each SISO nominal system. DBC can help the application of the H ∞ theory and the achievement of diagonal dominance. Application of DBC on a Bosch SCARA-type robot is presented to achieve decentralized joint control. Simulation results confirm the viability of ...