Closing the Loop with Liquid-Metal Sensing Skin for Autonomous Soft Robot Gripping

摘要

Soft robots are often limited in high-level decision making and feedback control due to a lack of multimodal sensing capabilities and material architectures that tightly integrate sensing and actuation. However, the recent development of elastic multimodal sensing skins has created the opportunity for closing the loop in soft robotic systems. In this work, we present a sensor-based finite state machine for a soft sensorized gripper mounted to a 4-degree-of-freedom robot arm. The soft gripper actuates between binary open and close states by activating shape-memory alloy springs, and contains proximity, pressure, and orientation sensors. The closed-loop control is demonstrated through scanning, grasping, and sorting tasks driven by sensor feedback. Using a time-of-flight distance sensor, the system can calculate the length, width, height, and center of mass of an object within a 60 mm x 25 mm workspace. With the time-of-flight distance sensor, pressure sensors, and inertial measurement unit, the system can detect and respond to external perturbations that interfere with the grasp, release, and transport of the object. The control strategy demonstrated in this paper can be expanded in the future to integrate basic autonomy in other soft robot systems.