Towards Robotic Feeding: Role of Haptics in Fork-Based Food Manipulation

Abstract

Autonomous feeding is challenging because it requires manipulation of food items with various compliance, sizes, and shapes. To understand how humans manipulate food items during feeding and to explore ways to adapt their strategies to robots, we collected a rich dataset of human trajectories by asking them to pick up food and feed it to a mannequin. From the analysis of the collected haptic and motion signals, we demonstrate that humans adapt their control policies to accommodate to the compliance and shape of the food item being acquired. We propose a taxonomy of manipulation strategies for feeding to highlight such policies. As a first step to generate compliance-dependent policies, we propose a set of classifiers for compliance-based food categorization from haptic and motion signals. We compare these human manipulation strategies with fixed position-control policies via a robot. Our analysis of success and failure cases of human and robot policies further highlights the importance of adapting the policy to the compliance of a food item.

Dimensionality Reduction and Motion Clustering During Activities of Daily Living: Three-, Four-, and Seven-Degree-of-Freedom Arm Movements

“…The application of this work is not task specific and is not exhaustive of the full set and complexity of motions within each task category, but instead provides a general framework that may be either applied in its current form for general use, improved on using fPCA, or could further be adapted to task specific scenarios to increase motion specificity. An example includes obtaining a partial hierarchy of motions exclusively for feeding [30]. The proposed approach could also be applied to a subset of the presented data, such as decoupling the reaching location from the wrist orientation.…”

“…Ultimately, verification of segmentation is performed heuristically by comparing results to predefined ground truth. Therefore, instead of implementing an automatic segmentation technique, we manually defined the start and end points of each motion segment by identifying when the end effector reached zero velocity, when a food item was acquired (analogous to [30]), completed a transfer or task, or returned the object to the table or the hand to its 'rest pose' (Table 1).…”

Upper-Limb Kinematic Parameter Estimation and Localization Using a Compliant Robotic Manipulator

“…4. The sequence of images at the top, which illustrates the experimental process, was recorded as a video 2 . The Cartesian trajectory in terms of position (graph at the top) and orientation (graph at the bottom) of x h G are also shown.…”

“…A SSISTIVE robots can be defined as devices equipped with sensory, perceptive, and cognitive capabilities to physically help disabled or elderly people in daily-life activities, thus circumventing the need for an attendant [1]. Typical existing applications involving this technology include activities that do not require high levels of physical interaction, such as feeding [2] or dressing [3]. However, the ability to facilitate tasks that involve high levels of physical interaction, such as climbing stairs or standing up from a bed or chair [4], relocating limbs [5], object handovers [6], and stabilizing limb motions [7], is also desired in assistive robots.…”

Towards Open and Expandable Cognitive AI Architectures for Large-Scale Multi-Agent Human-Robot Collaborative Learning

“…Regarding the operational scenarios where LfD has so far shown significant advances, these involve the use of both manipulator (e.g. manufacturing [32], assistive [33], healthcare [34], social [35], etc.) and mobile (e.g.…”