Comparative Study of Design of a 3-DOF Translational Parallel Manipulator with Prescribed Workspace

In this paper, a comparative study is performed between a deterministic and a robust optimization methods, for the design of a translational parallel robot with a prescribed workspace. The optimization is defined in each case as a multi-objective formulation problem. The deterministic formulation is proposed to maximize the robot’s compactness as well as its dexterity. For the robust formulation, it is proposed to combine two methods, i.e., Genetic Algorithms and Monte Carlo methods, where the previous two objective functions were optimized along with their two standard deviations. A sensitivity analysis is then performed on each of the cases to evaluate the impact of the uncertainties on the design parameters on robustness of the obtained optimal solution. These two optimization methods are applied to optimize a 3-DOF parallel manipulator called Tri-pyramid.