Decoupling control of flexure jointed hexapods.

摘要

As a member of the class of parallel manipulators, flexure jointed hexapods are great candidates for micro-precision applications in which only a very small workspace is required. This dissertation makes four major contributions to the research of flexure jointed hexapods. First, new decoupling algorithms are proposed. They exploit the properties of the joint space mass-inertia matrix of flexure jointed hexapods, loosen and remove the severe constraints imposed by previous methods on the allowable geometry, workspace, and payload. Second, a new identification algorithm is derived to estimate the joint space mass-inertia matrix, which plays a crucial role in the computation of decoupling transformations. The new identification algorithm, using an optimization criterion differing from the least squares criterion, also applies to a class of problems of estimating symmetric and positive definite matrices. Third, the relationships between different decoupling algorithms, disturbance rejection and robust stability is discussed. It is proven that optimal robustness can be achieved by choosing a unitary decoupling matrix. Finally, an approach for constructing optimal Jacobians for prioritized manipulations is described. A prioritized manipulation is a task in which some degrees of freedom (DOF) in the Cartesian space are more important than the rest. Thus the DOF can be divided into major DOF (MDOF) and secondary DOF (SDOF). Jacobians are constructed to achieve MDOFs while trading-off SDOFs with obstacle avoidance, faint tolerance, or joint motion optimization.