2025 Theses Doctoral

Robotic Exoskeletons to Assist Volitional Motor Control of the Paretic Hand

Chen, Ava

Recovering hand function after loss of voluntary movement, or paresis, from neurological injuries such as a stroke or spinal cord injury (SCI) requires the person to actively engage the impaired limb when practicing movement. However, reinforcing neuromotor coordination is difficult when the paretic hand does not respond to intent due to muscle weakness. In this dissertation, we build robotic exoskeletons that assist movement by acting on attempts to volitionally use the hand. We develop cable-driven mechanisms, control approaches, and evaluation methods to both study and leverage residual motor capacity after hand impairment.

The physical coupling of human and robot in wearable systems offers the potential to leverage body-powered movements for actuation and control. We use involuntary flexor synergies that are typical for stroke hemiparesis as the basis of stabilizing motion through antagonistic tension within active and passive cables in underactuated tendon-network transmission designs. We propose examples of such mechanisms for assisting thumb opposition and forearm supination. We use residual motor capacity for voluntary movements as the basis for interpreting intent and control over robotic exoskeletons. Individuals with C6-C7 SCI retain fine motor dexterity at the wrist, which we adapt for assisting grasp force modulation via a throttle-based control approach. Individuals with stroke hemiparesis experience difficulties in generating consistent muscle activation patterns when they do not have access to sensory feedback from movement. We propose a bidirectional training paradigm that mediates human-robot adaptation with augmented visual feedback of the model's intent predictions. These approaches for actuation and control enable intuitive, ipsilateral operation of robotic exoskeletons.

We seek insights from the experiences of clinicians and impaired users to better inform robotic design practice. Devices can achieve both rehabilitative and assistive aims when they enable the assessment, as well as execution, of intensive movement practice. We demonstrate the capabilities of a robotic hand exoskeleton as a sensing platform for making clinically relevant observations of the impaired body in motion. By continuously monitoring the change in finger resistance over multiple hand-opening cycles, we introduce the first observations of stiffness fluctuations of stroke-impaired fingers as a function of volitional movement. In a preliminary study, we propose cutaneous vibratory stimulation as a method for robotic devices to perform a dual role in assisting muscle tone and assisting volitional movement.

These projects advance multiple aspects of exoskeleton design and interaction that are needed to make devices adaptable to individual users, contexts of functional use, and degree of hand impairment. The process of developing devices to accommodate the specific needs of stroke and SCI populations takes an important step towards realizing user-driven, beneficial robots.