An observer-based inverse dynamics control strategy for flexible multi-link manipulators

This paper presents an observer-based inverse dynamics control strategy that results in small tip-position tracking errors for a class of multi-link structurally flexible manipulators. It was shown in Moallem et al. (1996) that small tracking errors are achievable by defining new outputs near the end points of an arm and augmenting the control inputs by terms that ensure stable operation of the closed-loop system. However, the control strategy required measurements of rates of change of flexible modes with time that are not conveniently measurable. A nonlinear observer is introduced here to estimate these variables by assuming that joint positions, rates and deflection modes are measured. By a proper choice of observer gains the error dynamics are guaranteed to consist of a stable linear part plus a bounded perturbation term that results in local asymptotically stable observation of deflection rates.