Optimization Design of a Walkable Fixture Mechanism

Abstract

In this paper, a novel fixture mechanism with combining a mobility of the legged robot and advantages of parallel mechanism is designed to hold the different size and shape, large-scale workpiece. The proposed mobile fixture mechanism holds the workpiece as a parallel manipulator while it walks as a legged robot. This kind of robotized fixtures can possess high self-configurable ability to accommodate a wider variety of products. In order to obtain the best kinematic dexterity and accuracy characteristics, comprehensive performance optimization is performed by non-dominated-genetic algorithm (NSGA-II). In the optimization procedure, a conventional kinematic transformation matrix (Jacobian matrix) and error propagation matrix are obtained through derivation and differential motion operations. The singular values and condition number based on velocity Jacobians and error amplification factors based on error propagation matrix are derived; in addition, relative pose error range of end effector is also derived. On the basis of the above measure indices, three kinds of nonlinear optimization problems are defined to obtain the optimal architecture parameters for better kinematic accuracy and dexterity in workspace. Comparison analyses of the optimized results are performed.