Stiffness optimization of a novel reconfigurable parallel kinematic manipulator

Abstract

This paper proposes a novel design of a reconfigurable parallel kinematic manipulator used for a machine tool. After investigating the displacement and inverse kinematics of the proposed manipulator, it is found that the parasitic motions along x-, y-, and θ z -axes can be eliminated. The system stiffness of the parallel manipulator is conducted. In order to locate the highest system stiffness, single and multiobjective optimizations are performed in terms of rotation angles in xand y-axes and translation displacement in z-axis. Finally, a case study of tool path planning is presented to demonstrate the application of stiffness mapping. Through this integrated design synthesis process, the system stiffness optimization is conducted with Genetic Algorithms. By optimizing the design variables including end-effector size, base platform size, the distance between base platform and middle moving platform, and the length of the active links, the system stiffness of the proposed parallel kinematic manipulator has been greatly improved.

A combinatorial search method for the quasi-static payload capacity of serial modular reconfigurable robots

“…First, a large number of computations are required to solve Eqs. (5) and (8) for multiple MRR configurations. In the stiffness analysis of traditional nonreconfigurable robots, the workspace of only one configuration is considered, which exhibits a complex problem already.…”

“…Since the deflection and torque objective functions in Eq. (5) are only required at one location on the assembled MRR, along with the lowest natural frequency, it becomes necessary to solve for the remaining global static deflections and natural frequencies throughout the entire MRR structure. The given objective functions are dependent on the remaining deflections, frequencies and unknown joint variables, which are greater in number than the system of equations in Eqs.…”

“…In the current literature, kineto-elastic analysis has been extensively studied for parallel reconfigurable robots [5,6]. However, there is a lack of research contributions for the kineto-elastic analysis of serial MRRs, which are more susceptible to adverse stiffness issues than their parallel counterparts.…”

“…Thus, direct evaluations of the deflections and torques are more reliable methods for assessing the stiffness performance of assembled MRR modules. For improving the kineto-elastic performance of robots, previous methods such as [5,6,[10][11][12] focus on implementing physical design changes on MRR modules, such as re-designing link geometry and adding stiffening springs to the joints. However, re-designing existing off-the-shelf modules is often a costly procedure, requiring extensive machining and modifications to the joint modules.…”

A New Modular, Autonomously Reconfigurable Manipulator Platform

“…These parallel manipulators have different task-properties than their serial counterparts, yet share the same principles of reconfigurability. A three-DOF reconfigurable parallel manipulator is presented in [11], in which an additional actuator changes the kinematics and dynamics of the manipulator, allowing for continuous and automatic reconfiguration of elements of the manipulator. A parallel manipulator named "Cheope" [12] is capable of altering its DOFs by transitioning into eight discrete configurations.…”