Advanced Tactile Sensing For Robotics

Author : Alex Makris Gruebele
Publisher :
ISBN 13 :
Total Pages : pages
Book Rating : 4.:/5 (124 download)

DOWNLOAD NOW!

Download or read book Robust Soft Sensors for Exploring the Real World written by Alex Makris Gruebele and published by . This book was released on 2021 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: While robots are gaining many abilities, they still lack the "magic of touch" to perform in contact-rich environments as humans do. Researchers working on touch for robots have disproportionately focused on developing tactile sensors for robot fingertips to improve robots' performance in grasping and manipulation tasks, and have made considerable progress, in part because they can leverage chips and circuit techniques developed for smart appliances and cell phones. However, like humans, robots also need to feel objects around them through more than just fingertip sensors to be able to negotiate and manipulate objects in complex environments. Think about how you might grab a plate off a surface, first grazing the back of your hand on the table as you reach under the rim, or how you guide your hand into your pocket or purse to fish out your car keys. This whole-hand aspect of tactile sensing has not advanced nearly as much as fingertip sensors have, one reason being that it requires covering large, curved, and often flexing surfaces. This capability was generally not possible prior to the explosion of soft robotics research in recent years. Prior approaches have often tried to use existing, essentially rigid, sensing elements and processing chips, embedding them on flexible circuits. But it is difficult to make something that is inherently two-dimensional and not stretchable conform to a flexing, three-dimensional skin. In addition, the sensors and chips are typically fragile with respect to impact, undermining the use case since contact -- whether intentional or accidental -- is the main point of a sensory skin. It's also expensive to build such a skin from many discrete elements and results in a tangle of wires, which must also accommodate stretching and flexing. I present a new approach that addresses these issues directly: taking the required rigid processing units and locating them away from the sensing site, where they are protected from impact and do not need to flex and stretch. I demonstrate this approach via three designs that illustrate different tradeoffs among sensor density, force resolution, and robustness. The first is a sensor for soft robots consisting of modular flexible bands that contain some rigid components. It protects them with a bending armor laminate and transmits signals through a flexible bus to a processing hub. The second is a stretchable capacitor-based covering with no rigid components that allows a robot both to pick up objects and feel its environment with sensors on the backs and sides of its fingers. The final design is completely conductor-free and thus extremely stretchable. It consists of silicone pouches with pneumatic channels to transmit sensor signals to remote pressure sensing chips. I describe the design challenges and new fabrication methods developed for each of these soft sensing technologies, along with results of experiments to characterize them and demonstrate their use in grasping and exploration. Together, the three approaches address new ways to cover robots with sensory capabilities at a range of scales -- from fingertip size -- to meters in length, and without having to design sensing components and wiring into the robot itself.