Tadesse, Yonas2023-08-302023-08-302022-052022-05-01May 2022https://hdl.handle.net/10735.1/9826Artificial hands with many unique designs and capabilities have been presented in the literature; however, most of these hands only have binary finger states, meaning only open or closed states. Position control of fingers for these hands enables more precise manipulation of objects for a wide variety of applications, expanding their usage beyond simple grasping tasks. An additional area of focus, particularly with thermally actuated artificial muscles, is increasing actuation frequency, which is currently limited by heat transfer rates. Improving frequency will allow for faster response of the hands to input stimuli and accelerate their potential usage in robotic and prosthetic applications. The objective of this research is to improve upon existing robotic hand designs that utilize thermally actuated artificial muscles. Such muscles have a high strength-to-weight ratio, low profile and silent actuation, but suffer from low actuation frequency and energy inefficiency. In this thesis, coiled shape memory alloy (SMA) muscles were studied and utilized for the design of a robotic hand using soft 3D-printed thermoplastic polyurethane (TPU). First, discussion on the design of the hand, manufacturing and characterization are presented. Next, finger position control to the hand using embedded flex sensors to create a closed-loop system is presented. This strategy maintains the ease of manufacturing of the TPU hand via 3D-printing, while also introducing new sensing capability. An investigation into improving the actuation frequency of the coiled SMAs using thermoelectric cooling via Peltier plates is also conducted. Holistically, these additions to the 3D-printed hand are projected to increase its overall actuation speed and precision, allowing for greater manipulation capabilities, but also adding additional weight and complexity to the overall system.application/pdfenEngineering, MechanicalThesis2023-08-30