Adaptive control for holonomically constrained robots: time-invariant and time-variant cases

This paper presents a general approach of model-based adaptive hybrid control for holonomically constrained manipulators. Both time-invariant and time-variant constraints are dealt with. This controller works on the basis of an orthogonalization principle of trajectory and force errors in the joint space and intrinsic characters of the manipulator dynamics. Its asymptotic stability is guaranteed by a passivity-based Lyapunov function. A distributed cooperation controller based on the same concept for multiple manipulators is also presented. Further, we show a parallel implementation of this controller using transputers on a six DOF direct drive manipulator, and some experimental results.