Assessing Control of Fixed-Support Balance Recovery in Wearable Lower-Limb Exoskeletons Using Multibody Dynamic Modelling

摘要

Despite many lower-limb exoskeletons requiring the use of crutches to maintain upright postures, limited research has assessed control of standing balance recovery in these systems. Using a model-based approach, the current simulation study investigated the performance of impedance controllers designed to assist with standing fixed-support balance recovery. A novel multibody dynamic model of the integrated humanexoskeleton system was designed to move in the sagittal plane. Development of the exoskeleton model (Technaid Exo-H3) was accompanied by parameter identification. The balancing torques produced by the human in the model were derived from offline linear control methods and saturated to approximate the torque-production of a young individual either with or without incomplete spinal cord injury. Without intervention, the injured user was not able to recover upright posture following a forward push of specific magnitude. Thus, three feedback control laws, inspired by robotics research (exoskeletons and humanoids), were implemented in the simulated exoskeleton. Each law assisted with balance recovery via reference tracking within the joint and/or whole-body center of mass space. Following optimization of control parameters, all proposed exoskeleton control laws were successful in assisting the injured user return to an upright posture. Joint space control yielded the best jointlevel reference tracking during recovery, while center of mass control better reduced forward center of mass excursions – albeit at the cost of joint-level tracking accuracy.