A new hybrid force/position control approach for time-varying constrained reconfigurable manipulators.

In this manuscript, a new hybrid force/position control approach has been proposed for time-varying constrained reconfigurable manipulators. In order to design the controller, firstly a reduced-order dynamic model of time-varying constrained manipulator system is presented. The uncertainties in the dynamical model of the system are inevitable; therefore the model-based control approach is inadequate to handle these systems. Therefore, inspired by this consideration, whatsoever partial information is available about the dynamics of the system, have been used for controller design purpose. The model-dependent control scheme is integrated with the neural network-based model-free control scheme. Radial basis function neural network is used for the estimation of the unknown dynamics of the system. Next, to overcome the aftereffects of the friction terms and neural network reconstruction error, an adaptive compensator is added to the part of the controller. For the stability analysis of the presented control scheme, the Lyapunov theorem and Barbalat's lemma are utilized. The designed control scheme guarantees that tracking errors of the joints and the force tracking error remain inside the desired levels and the joint tracking errors converge to zero asymptotically. Finally, comparative computer simulations show the superiority and the applicability of the developed control method applied over a 2-DOF time-varying constrained reconfigurable manipulator.