Assistive robot arms can help to provide care for those with disabilities. A latest paper on arXiv.org looks into robot-assisted feeding. A bite transfer strategy need to optimize its trajectories on the fly by bringing food stuff into a mouth devoid of sacrificing user comfort and ease.
In get to accomplish that, the researchers present a bite transfer algorithm for picking out trajectories in a ongoing place of mouth sizes, food stuff geometries, and poses. The method normally takes as input a food stuff mesh and an acquisition pose on the fork from the serious earth and generates an analogous simulation atmosphere. A constraint design is discovered to sample objective food stuff poses close to the mouth. Motion preparing is carried out to condition the perceived comfort and ease and bite volume efficiency of every transfer.
A user study shows that the proposed method of mixing comfort and ease and efficiency is the only approach to outperform a set pose baseline with statistical significance.
Robot-assisted feeding in domestic environments is hard because it needs robots to produce trajectories that properly deliver food stuff goods of varying shapes and sizes into the mouth when earning positive the user is relaxed. Our critical perception is that in get to remedy this challenge, robots need to balance the efficiency of feeding a food stuff item with the comfort and ease of every particular person bite. We formalize comfort and ease and efficiency as heuristics to integrate in motion preparing. We present an method dependent on heuristics-guided bi-directional Fast-checking out Random Trees (h-BiRRT) that selects bite transfer trajectories of arbitrary food stuff item geometries and shapes employing our created bite efficiency and comfort and ease heuristics and a discovered constraint design. True-robot evaluations exhibit that optimizing equally comfort and ease and efficiency significantly outperforms a set-pose dependent approach, and users desired our approach significantly much more than that of a approach that maximizes only user comfort and ease. Films and Appendices are observed on our internet site: this https URL.
Investigation paper: Belkhale, S., Gordon, E. K., Chen, Y., Srinivasa, S., Bhattacharjee, T., and Sadigh, D., “Balancing Effectiveness and Comfort in Robot-Assisted Bite Transfer”, 2021. Url: https://arxiv.org/abs/2111.11401