Soft Finger Grasp Force and Contact State Estimation from Tactile Sensors

TL;DR

This study investigates the efficacy of estimating finger force from integrated soft sensors and uses a neural network to estimate contact states, and applies this model in a plug-in task scenario and demonstrates its validity in estimating contact states.

Abstract

Soft robotic fingers can improve adaptability in grasping and manipulation, compensating for geometric variation in object or environmental contact, but today lack force capacity and fine dexterity. Integrated tactile sensors can provide grasp and task information which can improve dexterity,but should ideally not require object-specific training. The total force vector exerted by a finger provides general information to the internal grasp forces (e.g. for grasp stability) and, when summed over fingers, an estimate of the external force acting on the grasped object (e.g. for task-level control). In this study, we investigate the efficacy of estimating finger force from integrated soft sensors and use it to estimate contact states. We use a neural network for force regression, collecting labelled data with a force/torque sensor and a range of test objects. Subsequently, we apply this model in a plug-in task scenario and demonstrate its validity in estimating contact states.

Figures (13)

• Figure 1: Soft fingers (a) with integrated soft pressure and strain sensors grasping an object, and (b) example internal grasp states which can be detected based on the total finger force.
• Figure 2: Grasping force model, where each finger exerts a total force on the grasped object. The total forces can be measured by either the bulk deformation of the finger from an equilibrium position (right, top), or the total integral of surface pressure (right, bottom).
• Figure 3: Experimental setup for internal grasp state experiments
• Figure 4: We validate the ability to use internal forces for soft finger slip. The object size, robot offset and material affect the normal and cross-forces, and the stick, micro-slip and sliding can be distinguished. The results indicate (1) the Coulomb model provides 20% error, showing measuring normal force is useful for predicting slip, (2) the friction model also applies to two-finger applications.
• Figure 5: Soft gripper system with soft sensors (top) as well as characterization of sensors in simplified tasks (bottom).