Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 149843

149843 - Single-Chambered Soft Robot Actuation and Sensing Integration 

Within the field of robotics, soft robotics has recently emerged under a completely different paradigm than traditional robotics. Instead of inflexible materials that quickly and repeatably perform tasks with high precision, this sub-field aims to produce robots that easily deform and use this property to solve largely different problems than their hard counterparts and in different ways. The inherent deformability of soft robots promises solutions in navigating unstructured environments, collaborating safely with human counterparts, and manipulating fragile objects. This promise, however, has only partly been realized. Three key challenges that prevent soft robots from realizing their potential are actuation, sensing, and control. This work largely focuses on the first two. 

 

Thus far, actuation has mostly been solved using multi-chambered robots by which a soft robot can move via inflation of selective chambers, causing the robot to bend toward the contralateral side of that chamber. The inflation system usually requires numerous individual pneumatic tubes creating a tether between the robot and the air distribution network; limiting autonomy. Instead of relying on individual chamber inflation to drive the robot, we simplify the design to a single chamber with a fabric reinforcement. The stiff material of the fabric reinforcement provides a mechanically programmable trajectory for a robot although this limits the robot to a single executable trajectory. Our approach to extending available poses beyond those of a single trajectory is to infiltrate the walls of the tube with a phase change material (PCMs) which alters phases (i.e. fluid to solid or solid to fluid) via application of a stimulus. Thermally activated PCMs (wax) decrease viscosity with the application of heat; magnetically activated PCMs, (magnetorheological fluid) increase viscosity with increasing field strength. Directly integrating these actuation materials into the robot walls gives the robot some ability to deviate from the mechanically programmed trajectory invoked by the fabric reinforcement. This continuously varying family of curves constitutes an enlarged robot workspace [1].

 

Actuating the robot, however, is only part of the issue. The robot must also be able to sense the effect this actuation and any environmental disturbances have on its pose. Traditional robots use encoders at known joint locations connected by rigid links. Soft robots deform continuously across their entire body. This means that the robot must use measured strains to estimate its pose. Because of complex interactions with the environment, estimation requires large amounts of strain data. In addition to needing many sensors, the sensors should be highly stretchable and should also bear some of the mechanical load of deformation. This prevents their incorporation into the skin of the robot creating weak points. Optical waveguide sensors have both of these characteristics. These sensors guide light and lose more light as they are deformed. The light’s loss and travel time can be used to estimate how the guide is deformed. These fibers can be easily incorporated into the skin of a soft robot allowing estimation of the robot’s deformation including complex deformations like buckling [2].

 

In this work, we present an approach to resolving the integration of sensing and movement in a single-chambered soft robot. A precision-machined, metal vacuum mold allows accurate integration of both sensing and actuation materials into the thin walls of the robot. This mold also doesn’t create the failure initiation sites that 3D-printed ones do [3]. This process consists of a two-step molding process in which a very thin layer of elastomer is deposited over a metal core, and subsequently, the actuation and sensing materials are overlaid atop the cured layer. The second layer forms around the additions creating a final wall thickness of 2 mm, which encapsulates both sensing and actuation materials.

 

[1] Williamson JG, et al. “Extending the reach of single-chamber inflatable soft robots using Magnetorheological Fluids”

[2] Williamson JG, et al. “Stretchable optical waveguide sensor suitability for wrinkle degree detection in soft robots”

[3] Bui et al. “Endurance tests for a fabric reinforced inflatable soft actuator.”

Presenting Author: Caroline Schell University of Tusla

Presenting Author Biography: Caroline M. Schell is a Mechanical Engineering PhD student at the University of Tulsa. She has a Bachelor of Science in Petroleum Engineering and a Master's of Science in Mechanical Engineering from the University of Tulsa. As a material scientist, Caroline loves the inspection and creation of materials, finding uses for them and ideating new designs. Caroline's work focuses on the implementation of vascular networks in soft materials with an emphasis on phase changing fluids for tunable stiffness. When she's not working in the lab, she is playing videogames or knitting/crocheting.

Authors:

Caroline Schell University of Tusla
John Garrett Williamson University of Tulsa