Enhanced Force Estimation and Autonomous Force Tuning Through Fusion of Tactile Sensor and Quasi-Direct Drive Gripper

Abstract

Robotic manipulation relies on accurate force sensing and control mechanisms to ensure efficient and safe operation. This article presents a novel approach to enhance robotic manipulation capabilities by integrating a barometric tactile sensor with a quasi-direct drive gripper (QDDG). The synergy between these technologies addresses inherent limitations: QDDGs struggle with high current requirements and potential overheating, while tactile sensors often have restricted force ranges and sampling rates. By fusing data from these systems using an adaptive sensor fusion model, this study improves force estimation accuracy, extends force range capability, and boosts sampling rates up to 100 Hz. The integrated approach not only enhances grasping efficiency but also enables autonomous adjustment of grip force levels to prevent slippage. Experimental results demonstrate significant improvements in force estimation accuracy over individual sensor outputs, validating the effectiveness of the proposed integration for advanced robotic manipulation tasks. Furthermore, based on this sensor fusion framework, a slip detection and control mechanism was developed capable of grasping and lifting a diverse array of objects with high success-rates (>90%).