Design Theory And Methodology Of The TriVariant-A 5-DOF Reconfigurable PKM Module

This dissertation deals with the theory and methodology for the design of a 5-DOF reconfigurable hybrid robot module, including conceptual design, workspace synthesis and analysis, forward position analysis, rigid-body dynamic formulation and performance evaluation. The following contributions have been made. The theory of structure synthesis of sub-6 DOF parallel mechanisms is highlighted and a set of criteria for the conceptual design of reconfigurable parallel kinematic machine (PKM) modules is addressed. It concludes that the most suitable candidates for building reconfigurable PKM modules, would be the 3-DOF parallel mechanisms having a constrained limb whose type and degrees of freedom are the same as those of the mobile platform. As a result, a novel 5-DOF hybrid PKM module named TriVariant is proposed. It also demonstrates via examples that the TriVariant inherits all structural characteristics and reconfigurability of the well-known Tricept robot. The effect of the dimensional parameters on the kinematic performance of the TriVariant robot is investigated by means of monotonic analysis, leading to a set of appropriate constraints including the nearly axially symmetric manipulability and the maximum/minimum length ratio of the UPS limb, etc. Consequently, the optimal kinematic design of the TriVariant robot is implemented by solving a set of nonlinear algebraic equations. It shows that the TriVariant robot has competitive kinematic performance in comparison with the Tricept robot. Forward position analysis of the TriVariant robot is carried out using both analytical and numerical methods, compared with that of the Tricept robot. It shows that the forward position problem of the TriVariant robot is much simpler than that of the Tricept robot in terms of formulation and solution. Dynamic model of the TriVariant and the Tricept robots is formulated using the virtual work principle. Both local and global conditioning indices are proposed for the dynamic performance evaluation. It concludes that the TriVariant and the Tricept robots have similar dynamic performance provided that they share identical task workspace, and similar dimensional and inertial parameters. In addition, an effective approach to estimating the servomotor parameters of the TriVariant robot is presented using the singular value decomposition technique. The validity of this approach is justified by typical path planning. The outcome of this dissertation has provided a complete theoretical package for the development of the TriVariant robot, which is being used as a plug-and–play module to build a novel multi-functional device for steel girder welding and machining.