UCLA

Between a global pandemic, aging population, and labor shortages, there is an ongoing spike in the demand for healthcare that cannot be satisfied with traditional methods and the human workforce. Robotic technology offers a solution to this dilemma; applying robotics to healthcare is an active area of research that has begun to be widely commercialized. Whilethere are many potential avenues for robots to improve health and well-being, my research has focused on two areas in particular: the development of robotic hand exoskeletons for rehabilitation and the enhancement of robotic surgery via automation and sensor development. Exoskeletons have been shown to be effective for use in physical rehabilitation of numerous conditions including spinal cord injury and stroke. However, at present, the huge majority of exoskeleton systems are constructed for only the arm (from shoulder to wrist), back, lower limbs, or hands. Few systems have been developed that combine both full arm and hand systems, and those that do generally have limited actuation of the hand. This is partly because the mechanical complexity of the hand requires rigid hand exoskeletons to be complex and bulky if they are able to control many of the important degrees of freedom. This bulk and complexity makes the hand systems challenging to successfully integrate onto the distal end of an arm exoskeleton. However, there is significant demand for combined arm and hand rehabilitation exoskeletons because many activities of daily living, that physical therapy focuses on retraining, require reaching and grasping together. The overarching goals of this research are to develop a novel hand exoskeleton, experimentally evaluate its capabilities in preparation for application to stroke rehabilitation, and integrate it on the existing EXO-UL8 and concurrently developed BLUE SABINO upper limb exoskeleton systems. Chapter 1 provides an introduction on hand exoskeleton systems, with a focus on those designed for rehabilitation. My work on hand exoskeletons started by inheriting a hand exoskeleton mechanical design from Brando Dimapasoc, a graduating Master’s Degree student, that was developed as part of NSF Award #1532239. The system was intended to be attached to the EXO-UL8 and BLUE SABINO arm exoskeletons, had six active and six passive degrees of freedom, had three reconfigurable linkages to control the thumb and two groupings of fingers, used a bowden cable transmission system to enable remote placement of actuators, and was optimized to fit 90% of the general population. However, the design had only been tested as a 3D-printed prototype in a modified and simplified form. Further, the necessary electronic hardware (other than motor and sensor selection) and control software had not been started. Thus, the first stage of my research was to bring this first-generation hand exoskeleton to a functional state. This involved the mechanical assembly, the electrical design and assembly, and the software and control development of the system. Through testing, it was determined that a significant number of improvements must be made to the system in order for it to be suitable for use. Details of this work are contained in Chapter 2. With the lessons learned from development and testing of the first-generation hand exoskeleton, the next stage of my research involved the nearly complete redesign of the system in order to create the second-generation hand exoskeleton named the “Opposable-Thumb Hand Exoskeleton for Rehabilitation” or “OTHER Hand”. As the name implies, the system is designed to control opposition/reposition of the thumb in addition to the flexion/extension of each digit. This is a notable feature, not only because of the importance of opposition/reposition in many grasping tasks, but also because only a handful of exoskeletons in the literature control this motion. The OTHER Hand shares a number of features with the first generation system, though the execution of each is different. It attaches to both the EXO-UL8 and BLUE SABINO arm exoskeletons, is actuated using a Bowden Cable transmission such that the motor pack can be located remotely, has three reconfigurable linkages to control the thumb and two groupings of fingers in order to enable nearly all grasps, and is optimized to fit 90% of the population. The system has six active and eight passive degrees of freedom per hand. Chapter 3 documents the design of the OTHER Hand. Due to the numerous novel design choices made for the OTHER Hand, combined with the mechanical complexity of the hand in general, and thumb in particular, it is not feasible to know with certainty the types of grasps that can be actuated in the exoskeleton for the wide range of hand shapes and sizes. As such, it is necessary to validate the design of the OTHER Hand through testing with a group of subjects. This was accomplished through adaptation of the Anthropomorphic Hand Assessment Protocol for use with an exoskeleton to test the ability of thirteen subjects to grasp and manipulate 25 objects of the Yale-Carnegie Mellon-Berkeley Object Set using eight prehensile grasps and two non-prehensile hand postures. Additionally, the OTHER Hand was mounted on the EXO-UL8, and both systems were manually controlled to verify compatibility, workspace, and ability to bi-manually grasp a sample object. Chapter 4 presents the testing protocol and results. While exoskeletons for rehabilitation is an increasingly popular research area, robotic surgical platforms already have widespread commercial use and profound effects on clinical outcomes. Classically, these systems are controlled directly by a surgeon at a console in the same or adjacent room. They can augment the senses and movement precision of the surgeon during open or laparoscopic surgery in order to enhance the surgeon’s skills. However, surgeons commonly work exceptionally long hours in an environment where a single mistake can be fatal. Additionally, certain surgical subtasks are time-consuming, repetitive, and common to many different operations. Automating these subtasks has the potential to reduce the burden on surgeons while standardizing outcomes. Automation of one such subtask, soft tissue manipulation, is described in Chapter 5. Cataract surgery ranks among the most common operating room procedures worldwide. The aim of the surgery is to replace the clouded biological lens with a clear synthetic lens. Despite the prevalence, this operation is currently performed manually by a surgeon, and generally is fast, standardized, and safe. However, the human body is notably non-optimal for performing cataract surgery due to the transparency and fragility of the tissues of the eye. In order to remove the lens, it is standard to break it apart with phacoemulsification, use an irrigation/aspiration handpiece to aspirate the lens material, and then polish any remaining lens material from the capsular bag. Unfortunately, the back of this bag, the posterior capsule, is transparent, mere microns thin, and easily ruptured from contact, ultrasound energy, or pressure. Rupturing the posterior capsule causes the vitreous of the inner eye to spill out, resulting in critical failure of the surgery. Additional information about the location of the tool tip within the eye could be used to reduce the risk of such a failure. To this end, a proof-of-concept modification of a tool to add bioelectrical impedance sensing and tissue classification was developed and tested on porcine eyes. This research is summarized in Chapter 6.