Team Members & Role:
- SAVANNAH MACERO – LEAD & ANALYSIS
- RACHEL BERRY – LEAD & DESIGN
- NATALIJA BESLIC – ANALYSIS & EVALUATION
- MILLAN TARANTO – DESIGN & FABRICATION
- JOHNATHAN CZERNIK – FABRICATION & EVALUATION
Abstract:
Humans have a variety of ways of completing the same task, creating endless solutions for the same question within a short period of time. In the example of a human picking up an object and placing it somewhere else, one may ask why they would complete the task by deviating from the optimal trajectory. To investigate this phenomenon in a controlled environment that removes the complexity of human factors, a robot may be used. The UMass Human Robotic Systems Laboratory utilizes a Franka Emika robotic arm for research on learning by demonstration. The current configuration of the arm includes a 2D linear actuated gripper with prongs which can be manually manipulated by the human user to execute grasping tasks involving objects of varying sizes, shapes, and materials. This design lacks user-friendliness due to low back drivability of the prongs and a lack of a human hand interface. This results in unnatural movement, making it unsuitable for collecting substantial data. The inHand Gripper is designed to be used with a single human hand to train the robotic arm to perform grasping tasks with greater human-like movement. Its form factor closely aligns with the human hand, thereby maintaining user-friendliness with the naturalness of demonstrated movements. The rack and pinion design enables coupling of the prongs and smooth back drivability, ensuring the accurate demonstration of tasks. Moreover, the prongs of the gripper are designed to handle objects with varying shapes and materials, allowing for the effective grasping of all objects compatible with the original Franka Emika Gripper. Compared to other manual grippers currently on the market, which typically require significantly more funding, maintenance, and operating personnel, our design is streamlined and efficient. In the future, the resulting prototype or a derivative of it could serve as a valuable tool for collecting human data in studies on learning by demonstration.
