Haptic search with the Smart Suction Cup on adversarial objects

TL;DR

The paper addresses the fragility of vision-based suction grasp planning in adversarial or unseen object conditions. It introduces the Smart Suction Cup, a four-chamber, flow-based tactile sensing end-effector that localizes seal leaks without on-cup electronics and uses autonomous haptic search to refine grasp pose after contact. The method combines flow-derived lateral and rotational cues with axial force control to produce a model-free, yet principled, haptic controller, evaluated through CFD analysis and extensive bin-picking experiments that show up to a ~2.5x improvement over vision-only planning and robust performance on challenging objects. This approach improves gripping robustness in unstructured environments and offers a practical route toward integrating tactile feedback with existing vision-based planners in industrial automation.

Abstract

Suction cups are an important gripper type in industrial robot applications, and prior literature focuses on using vision-based planners to improve grasping success in these tasks. Vision-based planners can fail due to adversarial objects or lose generalizability for unseen scenarios, without retraining learned algorithms. We propose haptic exploration to improve suction cup grasping when visual grasp planners fail. We present the Smart Suction Cup, an end-effector that utilizes internal flow measurements for tactile sensing. We show that model-based haptic search methods, guided by these flow measurements, improve grasping success by up to 2.5x as compared with using only a vision planner during a bin-picking task. In characterizing the Smart Suction Cup on both geometric edges and curves, we find that flow rate can accurately predict the ideal motion direction even with large postural errors. The Smart Suction Cup includes no electronics on the cup itself, such that the design is easy to fabricate and haptic exploration does not damage the sensor. This work motivates the use of suction cups with autonomous haptic search capabilities in especially adversarial scenarios.

Figures (12)

• Figure 1: The multi-chamber Smart Suction Cup grips an adversarial object. The cup has four internal chambers, each connected to a pressure transducer that provides a measure of internal flow rate. It is able to localize small breaks in the seal due to, for example, the rugosity (e.g., wrinkles, bumps, etc.) of the object surface. Haptic search can allow for successful gripping even when the initial grasping point fails, important for visually-adversarial objects.
• Figure 2: Design of the end effector and the suction cup. (a) The end effector integration with the suction cup. (b) A close up of the suction cup shows how it is connected with a vacuum connector and hoses to the pressure sensors. (c) Cross-sectional view of the suction cup shows internal and outer dimensions.
• Figure 4: (a-b) Two cases of CFD simulation. Light yellow blocks are engaged objects and the cross-sectional view shows leak flow into channel number 1. (c-d) CFD result of the vacuum pressure measured at the sensor locations of each chamber. The bar graphs are from the maximum of the four vacuum pressures. (e-f) Cross-sectional view of the pressure distribution. The arrows inside represent the relative logarithmic scale of airflow velocity.
• Figure 5: System integration of the Smart Suction Cup. (a) the smart suction cup system integrated on UR-10 robotic arm with a 6 DOF F/T sensor and a microcontroller. (b) Close up of end-effector, including the depth camera.
• Figure 6: The reference frame associated with the tool end is shown, including the origin point ($O$) located relative to the unloaded cup lip. The cardinal directions of the cup are oriented along the walls of the inner chamber, shown in the bottom view.