In this paper, issues in the design of a hybrid passive/adaptive controller for a single flexible link robot with a payload mass are discussed. It is well known that collocated control of a single flexible link can be accomplished with PD control. It has already been shown that, with an appropriate choice of output variable (for example, the reflected tip position), a simple noncollocated "PD-like" controller, henceforth referred to as a passive controller, can provide good performance when the nominal physical parameters of the manipulator are known and do not vary significantly with time. In practical applications, a flexible link robot must be able to pick up and move unknown payloads. A fixed passive controller may not provide adequate tracking performance. In this paper, an adaptive control strategy is augmented to the passive controller introduced in to improve the tracking performance.

In this paper, issues in the design of passive controllers for flexible link robots are discussed. It is well known that, for the single flexible link manipulator, the mapping from the input torques to the joint velocities is passive whereas the mapping from the input torques to the derivative of net tip movements (the usual output chosen) is not. However, if an appropriate output variable is selected, the transfer function, can be made passive provided that the hub inertia seen by the flexible link is sufficiently small. Then, by the passivity theorem, using any strictly passive controller with finite gain will result in an L/sub 2/-stable system. However, for an an industrial robot, the effective hub inertia seen by the flexible link is usually large. Alternate approaches are examined for extending these passivity results to large hub inertias. Experimental results are presented for a 5-bar-linkage manipulator with the last link being flexible.>