Tactile Guided Control Strategy

Tactile guided control strategies aim to enable robots to perform dexterous manipulation tasks by integrating tactile sensing with control algorithms. Current research focuses on developing robust control methods, often employing model predictive control (MPC) or other learning-based approaches like neural networks (including graph neural networks) to process high-resolution tactile data from sensors such as GelSight and TacTip, and to achieve precise control in tasks such as grasping, surface following, and cable manipulation. This research is significant for advancing robotic dexterity and enabling robots to interact more effectively with the physical world in various applications, including manufacturing, household tasks, and human-robot collaboration.