Author(s) :

Sidharth Maheshwari .

Volume/Issue :

Abstract :

Robot-assisted feeding systems offer significant promise for individuals with partial motor impairment, yet delivering food comfortably to a human mouth remains a complex challenge. This work introduces a novel approach to motion planning in assistive feeding settings which both models and optimizes for human comfort. A comfort cost function was developed that integrates trajectory smoothness, velocity, and a novel “closeness” metric grounded in proxemics. This cost function serves as a guide to a new motion planning sampling strategy and a cost-aware shortcut smoothing for post processing. Through simulation experiments in PyBullet across various realistic assistive feeding scenarios, the proposed methods generated paths with 17% decrease in comfort cost compared to standard planners making it more suited for adapting to human comfort. While the approach shows substantial gains in environments with a wider free configuration space, improvements were limited in environments where the free configuration space was narrower. These findings suggest a promising foundation for comfort-aware motion planning, with implications for improving autonomy and independence in robotic caregiving.