Learning Fingertip Force to Grasp Deformable Objects for Soft Wearable Robotic Glove With TSM

Abstract

Soft wearable robotic gloves based on tendon-sheath mechanism are widely developed for assisting people with a loss of hand mobility. For these robots, knowing the fingertip forces applied to deformable objects is crucial in successfully grasping them without causing excessive deformations. Existing studies presented methods to predict fingertip force applied to rigid objects only using information from the actuation system. However, forces applied to deformable objects are subject to non-linearity and hysteresis in relation to the objects' stiffness, which further complicates the problem. Therefore, this letter proposes a deep-learning model that can accurately estimate the fingertip forces applied to deformable objects using motor encoder values, motor current, and wire tension. Our model is based on an integrated system of Long Short-Term Memory models that 1) estimates stiffness of the grasped objects and 2) incorporates the estimated stiffness for predicting the fingertip forces. When evaluated using a TSM-based soft wearable robot, the proposed model recorded fingertip force estimation of 0.702 N RMSE, achieving 45% increase in accuracy compared to LSTM that does not consider the objects' stiffness. The applicability of our method was evaluated by estimating the fingertip forces applied to common daily items and performing real-time force control.