Novel control law for predictor-based MRAC for overactuated systems

摘要

In this paper, novel control and update laws for predictor-based Model Reference Adaptive Control for the case of over-actuated systems are presented. A weaker assumption than positive definiteness of the control effectiveness matrix is formulated. This allows to take actuator failures and even changes in the control direction into account. Furthermore, it is shown that the number of adaptive parameters can be reduced avoiding over-parameterization. The control law is applied to a hexarotor example and simulation results demonstrate the control law's capabilities.