Novel two-stage uncertainty optimization design of a compliant finger based on stochastic perturbation approach

Abstract

This paper presents a novel design procedure of the finger of a compliant gripper, which is used for grasping objects of different shapes under uncertain loads. The compliant finger is divided into two stages for design: finger root and fingertip. The finger root is designed using compliant mechanisms of topology optimization. The stochastic perturbation approach is introduced for the first time into a compliant mechanism topology optimization model under load uncertainty to quantify the uncertainty of load and node displacement of the finger root. The objective function of finger root optimization is maximizing the expectation of the output displacement while minimizing its standard deviation. The fingertip is designed using continuum structure topology optimization based on stiffness optimization. The optimization problems are solved using the method of moving asymptotes. Then, three numerical cases show that, compared with deterministic optimization, the compliant finger designed considering load uncertainty exhibits reduced sensitivity to load direction uncertainty and possesses greater stiffness. The finger prototype is manufactured using 3D printing, and the maximum gripping weight of the compliant gripper is tested. The grasping experiments indicate that the designed gripper can grasp various objects' shapes and avoid damage to fragile and high-precision surface objects.