As an undergraduate engineering pupil in Delhi, India, Amisha Bhaskar took a area journey to a facility for disabled battle veterans and met a person who had misplaced each fingers. When she requested him what applied sciences may enhance his life, his reply left an indelible impression: He wished one thing so he may maintain himself and never be pressured to depend on others.
Now a second-year doctoral pupil on the College of Maryland learning laptop science, Bhaskar has centered on the wounded veteran’s broad request as her space of examine. Working with others within the Robotics Algorithms & Autonomous Methods Lab, she is growing an modern robotic instrument to assist individuals with mobility impairments feed themselves.
The staff’s work was acknowledged final month on the IEEE Worldwide Convention on Robotics and Automation (ICRA 2024) in Yokohama, Japan, the place a paper Bhaskar offered as lead co-author obtained prime honors in a specialised workshop on cooking and robotics. It’s out there on the arXiv preprint server.
Current robotic-assisted feeding know-how could be very restricted, the UMD researchers mentioned. Business robotic arms have a hard and fast, pre-programmed movement that enables them to select up meals solely in a particular spot on a plate, and so they lack the power to detect whether or not they’ve achieved that job.
“They don’t seem to be studying on the go, so it should simply preserve doing this movement irrespective of if you wish to eat it or not, or if the meals is picked up or not,” mentioned Bhaskar.
Robotic-assisted feeding may be divided into two steps, she defined: the “acquisition” step entails a utensil selecting up the meals, whereas the switch step is the method of the meals reaching an individual’s mouth with out being dropped or succumbing to another mishap.
Bhaskar and the UMD staff are presently engaged on the acquisition step, with a lofty aim. Whereas different analysis teams typically depend selecting up meals on a utensil simply as soon as as successful, the UMD staff’s goal is to clear the plate.
The system should have the ability to acknowledge and transport quite a lot of meals served in assisted-care settings—from liquid meals to semi-solid ones like yogurt and tofu to cereals.
One of the vital vital challenges for a robotic is dealing with meals with diverse textures and consistencies inside a single dish, the researchers mentioned. Ramen, for instance, presents a fancy state of affairs that features a liquid broth, squishy tofu, stable greens and irregularly formed noodles that stay the most important problem, Bhaskar mentioned. “Each single component requires completely different methods, a few of which must be mixed,” she mentioned
An interdisciplinary method has performed a key position within the challenge’s success, mentioned Pratap Tokekar, an affiliate professor of laptop science with an appointment within the College of Maryland Institute for Superior Laptop Research.
“The know-how we’re engaged on entails laptop imaginative and prescient, synthetic intelligence, deep neural networks, mechanical engineering and extra—all of it wants to come back collectively seamlessly in order that the robotic system is each protected for customers and environment friendly in undertaking the duty at hand,” he mentioned.
Tokekar is educational adviser to Bhaskar and one other graduate pupil engaged on the challenge, Rui Liu, a third-year doctoral pupil in laptop science.
Robotic-assisted feeding is a comparatively new space of analysis for Liu, who had beforehand centered on laptop imaginative and prescient and human-robot interplay. However like Bhaskar, Liu sees the potential right here to tremendously enhance individuals’s lives, significantly older adults or these with mobility points that make feeding themselves tough.
Further staff members embrace Vishnu D. Sharma, Ph.D. ’24 and Guangyao Shi, Ph.D. ’23, now a postdoctoral researcher on the College of Southern California.
Whereas the challenge might be a number of years away from real-world utility, Tokekar is assured within the staff’s progress, and significantly in Bhasker’s and Liu’s eagerness and intense focus.
“The very best a part of this challenge is that each time we meet, they’ve 10 new concepts because the final time that we met,” Tokekar mentioned. “As an alternative of me telling them what to do, they already know what to do. I am simply serving to form their concepts.”
Extra data:
Amisha Bhaskar et al, LAVA: Lengthy-horizon Visible Motion primarily based Meals Acquisition, arXiv (2024). DOI: 10.48550/arxiv.2403.12876
arXiv
College of Maryland
Quotation:
Researchers’ robotic system goals to enhance autonomy for individuals with mobility points (2024, July 1)
retrieved 1 July 2024
from https://techxplore.com/information/2024-07-robotic-aims-autonomy-people-mobility.html
This doc is topic to copyright. Other than any truthful dealing for the aim of personal examine or analysis, no
half could also be reproduced with out the written permission. The content material is offered for data functions solely.