This paper details the design and fabrication process of a fully integrated soft humanoid robotic hand with five finger that integrate an embedded shape memory alloy (SMA) actuator and a piezoelectric transducer (PZT) flexure sensor. Several challenges including precise control of the SMA actuator, improving power efficiency, and reducing actuation current and response time have been addressed. First, a Ni-Ti SMA strip is pretrained to a circular shape. Second, it is wrapped with a Ni-Cr resistance wire that is coated with thermally conductive and electrically isolating material. This design significantly reduces actuation current, improves circuit efficiency, and hence reduces response time and increases power efficiency. Third, an antagonistic SMA strip is used to improve the shape recovery rate. Fourth, the SMA actuator, the recovery SMA strip, and a flexure sensor are inserted into a 3D printed mold which is filled with silicon rubber materials. The flexure sensor feeds back the finger shape for precise control. Fifth, a demolding process yields a fully integrated multifunctional soft robotic finger. We also fabricated a hand assembled with five fingers and a palm. We measured its performance and specifications with experiments. We demonstrated its capability of grasping various kinds of regular or irregular objects. The soft robotic hand is very robust and has a large compliance, which makes it ideal for use in an unstructured environment. It is inherently safe to human operators as it can withstand large impacts and unintended contacts without causing any injury to human operators or damage to the environment.

Bistable mechanisms have two stable equilibrium positions separated by a higher energy unstable equilibrium position. They are well suited for microswitches, microrelays, and many other macro-and micro-applications. This paper discusses a bistable buckled beam actuated by a moment input. A theoretical model is developed for predicting the necessary input moment. A novel experimental test setup was created for experimental verification of the model. The results show that the theoretical model is able to predict the maximum necessary input moment within 2.53%. This theoretical model provides a guideline to design bistable compliant mechanisms and actuators. It is also a computational tool to size the dimensions of buckled beams for actuating a specific mechanism.

No abstract