Functional Eigen-Grasping Using Approach Heatmaps

TL;DR

This paper tackles functional grasping of everyday tools by introducing an approach heatmap that encodes optimal palm-approach regions tailored to the tool’s functional parts. A two-stage method first offline generates the heatmap by sampling reachable palm configurations and scoring them with directional manipulability, then runtime grasp planning uses eigengrasp with a Hybrid energy objective $E=\alpha E_c + \beta E_f + \gamma E_p$ to achieve functional grasps. Key contributions include the novel heatmap representation, compatibility with both anthropomorphic and non-anthropomorphic hands, and three energy terms that encode functional and reachability constraints. Experimental results show effective functional grasps on various objects and real-world validation, though limitations exist in gripper force and object geometries; future work aims to automate heatmap generation for unseen objects and integrate the heatmap with planning for better generalization.

Abstract

This work presents a framework for a robot with a multi-fingered hand to freely utilize daily tools, including functional parts like buttons and triggers. An approach heatmap is generated by selecting a functional finger, indicating optimal palm positions on the object's surface that enable the functional finger to contact the tool's functional part. Once the palm position is identified through the heatmap, achieving the functional grasp becomes a straightforward process where the fingers stably grasp the object with low-dimensional inputs using the eigengrasp. As our approach does not need human demonstrations, it can easily adapt to various sizes and designs, extending its applicability to different objects. In our approach, we use directional manipulability to obtain the approach heatmap. In addition, we add two kinds of energy functions, i.e., palm energy and functional energy functions, to realize the eigengrasp. Using this method, each robotic gripper can autonomously identify its optimal workspace for functional grasping, extending its applicability to non-anthropomorphic robotic hands. We show that several daily tools like spray, drill, and remotes can be efficiently used by not only an anthropomorphic Shadow hand but also a non-anthropomorphic Barrett hand.

Figures (9)

• Figure 1: Functional grasp. With the self-generated heatmap on the left and the functional part of the object as input, the grasp planner can generate a grasp capable of satisfying the intended function.
• Figure 2: Method Overview - Setting one of the fingers as the functional finger, the approach heatmap is generated in (A) Then we use it in the the grasp planning process using the eigengrasp, applying the energy terms specified in (B) to achieve the functional grasp.
• Figure 3: Simplified hand models used for sampling reachability of the object
• Figure 4: The manipulability ellipsoid of the index finger at different reachable points of the object
• Figure 5: Minimizing the contact energy is done by minimizing the dot product of n and o while also minimizing the distance o