Abstract
This research aims to develop the force control environment for a powered upper-limb exoskeleton dedicated to stroke assessment and rehabilitation. Extensive clinical studies and neuroscience experiments have shown that exoskeletons can enhance stroke recovery by promoting patient engagement through 1) delivery of therapy with increased duration and intensity, and 2) providing training with adaptive assistance based on patient performance. Furthermore, there is growing demand to automate aspects of the assessment and rehabilitation process, and to provide researchers with instruments to gain insights into the intricacies of arm and hand function.To address these needs, this dissertation contributes to the development of BLUE SABINO (BiLateral Upper-extremity Exoskeleton for Simultaneous Assessment of Biomechanical and Neuromuscular Output). This groundbreaking instrument combines a highly precise and repeatable dual-arm exoskeleton with an advanced electroencephalographic (EEG) and electromyographic (EMG) data collection system. BLUE SABINO aims to provide a holistic approach to quantitative assessment and evaluation of arm function, catering to both healthy and impaired users.
The preliminary design for BLUE SABINO was proposed prior to this work; however, specific improvements are necessary to achieve the required agility to investigate natural human upper limb motion. Firstly, an advanced electromechanical structure is required to support actuation and sensing of human inputs. Additionally, a human-robot attachment (HRA) system is needed to connect the user and exoskeleton physically. Finally, a force control scheme is needed to enable transparent physical human-robot interaction (pHRI) so that it is possible to observe the unperturbed arm function of the user.
This dissertation presents several improvements made to BLUE SABINO's design. These include the addition of mechanical limits to enhance user safety by preventing overextension, the development of a novel size adjustable HRA system that minimizes misalignment and accommodates 95% of human anthropometric sizes, and the creation of robust and highly mobile interior/exterior electronics wiring/housings to support the electromechanical components of the exoskeleton.
Furthermore, this dissertation also describes the development of a transparent force control scheme for BLUE SABINO. This involves the development of kinematic and force mappings of human-applied forces to joint torques and an admittance-control scheme with a novel velocity-state-error-damping term. This control scheme improves overall stability while reducing the required user effort, with high spatial/temporal accuracy compared to traditional admittance models.
Overall, this work advances the field by addressing robotic stroke assessment and rehabilitation challenges. The improvements made to BLUE SABINO's design and the transparent force control scheme enhance the safety, flexibility, and functionality of the exoskeleton, providing a promising platform for future developments in this domain.