Motion control of brachiation robot by using final-state control for parameter-varying systems

In this paper, we apply an error learning method of final-state control to motion control of a 3-link brachiation robot of which the root joint has no drive unit. A brachiation robot moves under a ladder across the ceiling by holding the ladder with arms (just like a monkey moves from a tree to another). We treat the nonlinear terms of the governing equation as if they were linear and represent the dynamic model as a linear parameter-varying system. By simulations and experiments it is demonstrated that the feedforward input obtained by the error learning method can bring the brachiation robot from an initial state to a desired state in a specific time.