🤖 AI Summary
Robotic manipulation of heterogeneous objects faces a fundamental trade-off between adaptability and precision, arising from diverse shapes, sizes, and fragility. Conventional rigid grippers achieve high precision but lack generalization, while traditional soft surfaces require dense actuator arrays, incurring high complexity and cost. This work proposes a vertically compliant, sparsely actuated manipulation surface: linear actuators are placed only at the perimeter with uniform spacing, supporting a continuous elastic skin that deforms cooperatively to enable full-surface adaptive manipulation. We introduce the first “sparse-actuation + continuous-soft-surface” architecture, achieving stable handling of delicate micro-objects—e.g., 3-mm-diameter strawberries and tofu—even when actuator spacing reaches five times the object size. Compared to conventional soft surfaces, our design reduces actuator count by 60%, cuts degrees of freedom to one-quarter, and triples operational area coverage—simultaneously delivering high adaptability, low dimensionality, and cost efficiency.
📝 Abstract
Object manipulation in robotics faces challenges due to diverse object shapes, sizes, and fragility. Gripper-based methods offer precision and low degrees of freedom (DOF) but the gripper limits the kind of objects to grasp. On the other hand, surface-based approaches provide flexibility for handling fragile and heterogeneous objects but require numerous actuators, increasing complexity. We propose new manipulation hardware that utilizes equally spaced linear actuators placed vertically and connected by a soft surface. In this setup, object manipulation occurs on the soft surface through coordinated movements of the surrounding actuators. This approach requires fewer actuators to cover a large manipulation area, offering a cost-effective solution with a lower DOF compared to dense actuator arrays. It also effectively handles heterogeneous objects of varying shapes and weights, even when they are significantly smaller than the distance between actuators. This method is particularly suitable for managing highly fragile objects in the food industry.