Defense: Novel Approaches to Soft Robot Actuation and Sensing; A Bioinspired Soft-rigid Hybrid Finger with Variable Stiffness Modules

Keng-Yu Lin
Computer Engineering PhD Candidate
Location
Virtual Event
Advisor
Michael Wehner

Join us on Zoom: https://ucsc.zoom.us/j/96351817704?pwd=RllJb21aRUFneTV1Q3F1RDlUeDBLUT09 / Passcode: 591725

Description: Soft robots are a new field in robotics that shows an increasing potential to dramatically expand the capabilities of robots. We have analyzed the advantages and disadvantages of traditional rigid robots and soft robots, as well as the timing and occasions for their application. The characteristics of these two types of robots are completely different. However, in real human life, we are faced with a situation that falls between the characteristics of these two types of robots. Sometimes we need to grasp fragile objects, sometimes we need to lift heavy objects, and sometimes we need to press buttons or electrical switches and other daily actions. We need to think more formally about whether soft robots can perform these tasks and what critical areas are still lacking that need to be addressed. Thus, their underlying subsystems (actuation, sensing, control) and their role in robotics must be reconsidered. We first rethink sensing, in which traditional rigid robot sensing is very intuitive, using as many sensors as their degrees of freedom to describe the state of the robot. However, for soft robots, sensing becomes very difficult. Secondly, we rethink actuation; soft actuators are great for applying a distributed force on fragile objects. But the backdrivability that gives them the advantage also limits their ability to generate high forces.

In this dissertation, we rethought the soft systems and proposed the fundamental novel types of soft sensor and actuator necessary to develop the field of soft robotics from interesting concepts to useful devices. We demonstrate a network of fiber-based displacement sensors to measure robot state (bend, twist, elongation) and two microfluidic pressure sensors to measure overall and local pressures. The fiber-based sensors are fundamentally designed to be used in groups and leverage the concepts from beam theory and mechanics of materials to infer system state from a strategically located system of sensors. Intended to be built into a soft robot at the system level, a properly configured array of these deformation and pressure sensors can give state awareness far beyond that of individual sensors.

We present a bioinspired soft finger with a soft-rigid hybrid structure that can have multiple curves and force direction controllability to provide force in a specific direction. The soft and rigid states of multiple independent locking modules can be controlled independently by connecting them into a chain-like system. We have performed a modeling analysis of this controlled soft-rigid module, which provides an adaptable and scalable design framework for future bioinspired robotic fingers. We propose a new soft-rigid hybrid structure design and a method of interaction between pneumatic and tendon-driven actuators. The two actuation methods can increase the finger’s flexibility, allowing it to grasp objects of various shapes, sizes, and weights quickly and stably while providing sufficient force in specific directions. We have also added a nail mechanism to the tip of the finger, which helps the finger grip flat or small objects.

Lastly, we studied the possible power source of soft robots and found that if we look at the energy density alone, chemical reactions can provide a higher energy source than those energy sources with rigid components. However, the challenge of chemical reactions is to control the fluid efficiently. We develop a normally open passive microfluidic valve with reduced-order control for a micro-combustion chamber. This passive microfluidic valve can be installed on a micro-combustion chamber and is responsible for all fluid control, including intake and exhaust. This novel passive microfluidic valve may also be used in other applications, such as sensors for soft robotics.