Physiologically-inspired exoskeleton controller to enhance human balance - PROJECT SUMMARY Reducing the prevalence and burden of falls and fall-related injuries is a significant public health concern. Lower limb robotic exoskeletons have the potential to augment human balance and prevent falls. However, current exoskeletons are commonly designed to improve walking, not reduce the risk of falling. For the potential of exoskeletons for enhancing human balance to be realized, they must provide appropriate assistance across the diverse range of conditions that elicit falls. However, this remains an open challenge. This proposal uses physiological inspiration to develop a generalizable framework for symbiotic exoskeleton control that enhances balance during standing and walking. Reactive balance control requires a coordinated response at both the hip and ankle. However, current exoskeletons are commonly single-joint exoskeletons, and use local control principles that ignore the coordinated response. To overcome this limitation, I will leverage Dr. Ting's expertise in reactive balance control that demonstrates that the coordinated multi- joint reactive balancing response is mediated by global feedback of center of mass (CoM) kinematics rather than local feedback from individual joints. This global CoM feedback strategy explains the physiological coordinated multi-joint balancing response to postural perturbations of varying directions and magnitudes during both standing and walking. Thus, using a global CoM feedback model to control a multi-joint exoskeleton has the potential to greatly simplify its control and generalize across different tasks and perturbation events. I will test this concept in Aims 1 & 2 using state-of-the-art multi- joint exoskeleton systems currently used in Dr. Sawicki and Young's laboratories. Specifically, I will evaluate the efficacy of a global CoM controller at enhancing reactive balance control during both standing and walking compared to a local joint controller. Lastly, to facilitate the future usage of our novel balance-augmenting exoskeleton framework outside the laboratory, in Aim 3, I will use data from a wearable sensor suite worn by the participant to estimate CoM kinematics—the input into the global CoM controller—in real-time, rather than relying on high-fidelity, lab-based measures. To achieve this, I will leverage recent advances in deep learning by Dr. Young that enable highly accurate real-time estimates of physiological states across a range of walking conditions that generalize to unseen gait conditions from a few worn sensors. Together, this work provides the framework for developing exoskeletons aimed at improving balance and reducing the risk of falling in high-risk individuals, including older adults, those with musculoskeletal injuries, or stroke survivors.
