Abstract
Solid Freeform Fabrication (SFF) techniques allow the in-situ fabrication of fully-assembled devices with mating/fitting parts. Recently, this technique of fabrication has been found to be useful for building integrated mechanisms in robotics, and a wide array of other similar applications are anticipated. An interesting issue that arises during the fabrication of such mechanisms is the determination of an optimal pose in which the mechanism should be built. For example, should the mechanism be built in a folded or stretched-out position? In conventional manufacturing these issues do not arise, as each individual link is typically manufactured separately and then the pieces are brought together during assembly.
In this paper, we address the issue of finding a preferred (or optimal) pose for in-situ fabrication of planar mechanisms. There are many factors (e.g. achievable tolerances, non-interference, workspace size limitations, thermal considerations etc.) which can determine the suitability of a candidate build pose so that pre-specified task requirements are met. We limit our analysis to finding the optimal build configuration given achievable (in general, non-homogeneous, anisotropic) accuracy on joint position. For this treatment, we also make the simplifying assumption that the task requirements can be best satisfied by minimizing variability of link-lengths. Alternate task requirements, for example, maintaining end-point accuracy within a tolerance region, are being considered as part of ongoing work. We cast the problem of minimizing variability in link length as that of determining the relative position of two location tolerance regions for which the difference between their extremal distances is at a minimum (i.e., as they undergo constrained relative motion in the Euclidean plane). The method is similar to computational geometry techniques that have been developed in pattern matching and robot motion planning, with some important differences. We present some example mechanisms and their optimal pose under given workspace configurations.
