Abstract

A soft robot is made from deformable and flexible materials such as silicone, rubber, polymers, etc. Soft robotics is a rapidly evolving field where the human-robot-interaction and bio-inspired design align. The physical characteristics such as highly deformable material and dexterity make soft robots widely applicable. A soft robotic gripper is a robotic hand that acts like a human hand and grasps any object. The most common applications of soft robotics grippers are gripping and locomotion in sensitive applications where high dynamic and sensitivity are essential. Nowadays, soft robotics grippers are used without any sensing method and feedback as it is crucial to make the output feedback from the gripper. The major drawback of soft robotic grippers is their need for more precision sensing. In traditional robots, we can integrate any sensor to detect the force and orientation of objects. Still, soft robotic grippers rely on the deformation sensing method, where the sensor must be highly flexible and deformable. With a precise sensing method, it is easier to determine the exact position or orientation of the object being gripped, and it limits the application of the soft robotic gripper. Sometimes, soft robots are employed in harsh environments to solve problems. With the sensing feedback, automation may become more reliable and succeed altogether. So, in this research, we have designed and fabricated a soft sensor to integrate with the gripper and to observe the feedback of the gripper. We propose integrated multimodal sensing that incorporates applied pressure and resistance change. The sensor provides feedback when the grippers hold any object, and the output response is the resistance change of the sensor. The liquid metal is susceptible and can respond to low force levels. We presented the 3D design, FEM simulation, fabrication, and integration of the gripper and sensor, and by showing both simulation and experimental results, the gripper is validated for real-time application. FEM simulation simulates behavior, optimizing design and predicting performance. We have designed and fabricated a soft sensor that yields microfluidic channel arrays embedded with liquid metal Galinstan alloy and a soft robotic gripper hand. Different printing processes and characterization results are presented for the sensor and actuator. The fabrication process of the gripper and sensor is adequately described. The gripper output characteristics are tested for bending angle, load test, elongation, and object holding under various applied pressure. Additionally, the sensor was tested for stretchability, linearity and durability, and human gesture integration with the finger, and this sensor can be easily reused/ reproduced. Furthermore, the sensor exhibits good sensitivity concerning different pressure and grasping various objects. Finally, we collected data using this sensor-integrated gripper and trained the dataset using machine learning models for automation. With more data, this can be an autonomous gripper with intelligent sensing methodologies. Moreover, this proposed stretchable sensor can be integrated into any existing gripper for innovative real-time applications.

Date of publication

Spring 4-10-2023

Document Type

Thesis

Language

english

Persistent identifier

http://hdl.handle.net/10950/4189

Degree

Master of Science in Electrical Engineering

COinS