CAREER: Robotic Augmentation of Human Reflexes and Reach through Collaborative Grasping
University Of California-Berkeley, Berkeley CA
Investigators
Abstract
This Faculty Early Career Development Program (CAREER) project will create a new generation of effective and affordable wearable robotic assistance for people with severe weakness or partial paralysis of the hands and arms, such as might arise from cervical spinal cord injury. The objective is to help the user with grasping and holding actions, and enable performance of tasks of daily living. The project focuses on three assistance methods. The first is to use the high reaction speed of the robot to anticipate and augment the user's own grasp reflexes. The second is to provide natural and comfortable collaborative grasping and holding, by having the robot adapt its response to user preference. The third combines both anticipatory reflex augmentation and adaptive collaborative grasping to bestow maximal benefits under conditions of realistic variability and uncertainty. This work addresses fundamental issues of shared human-robot collaboration and the design of assistive mechanisms. It advances the health and welfare of people with disabilities, and increases their ability to function independently. Other benefits to society are the development of new educational techniques to broaden participation of people with disabilities in engineering and science, both in the classroom and in the local community. For people with disabilities capable of partial body-power, miscoordination with a robotic exoskeleton can lead to inefficiency, slowness, and frustration. This project introduces an unexplored modality for assistive grasping devices called collaborative grasping, or co-grasping, which is suitable to augment people with partial body power. This approach fundamentally alters the role of robotic elements with the goal of improving shared task performance and device embodiment. The project pursues two primary functional outcomes: (R1) reducing the rate of accidental object drops using tactile sensors and robot-led reflexes and (R2) easing body movements with robotic support to improve comfort and skill during reach-and-grasp. The methods used include the modification of a novel co-grasping testbed to also enable human-only and robot-only control cases, to probe interwoven questions of intelligence and embodiment through human subjects experiments. Specific experiments investigate the role of inherent haptic feedback on physical human robot interaction and user perception during challenging grasping tasks. The capstone of this project (R1+R2) is a collaborative arbitration prototype to accomplish (R1) and (R2) concurrently, including during realistic tasks of daily living. Overall, this work generates new tools and knowledge to improve and personalize the integration of robotic systems with the hands to enhance human independence and wellbeing. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
View original record on NSF Award Search →