A multiestimation-based scheme for adaptive control of robotic manipulators guaranteeing closed-loop stability

A multiestimation-based adaptive controller is designed for robotic manipulators. The scheme is composed by a set of estimation algorithms running in parallel. A supervisory switching logic decides the estimation algorithm which parameterizes the adaptive controller in real time according to an identification performance index. Multiestimation based techniques can improve the transient response of adaptive systems by appropriate switching between estimation schemes as it has been verified in simulation examples. A minimum residence time between consecutive switching is stated in order to prove closed loop system stability. The robot is controlled by using an impedance type control. Furthermore, a correcting term has been added to the standard control law used in the literature. The incorporation of this term allows choosing independently the estimation algorithm from the modified control law what leads to a degree of freedom in the choice of the estimation algorithm.