Modeling and control of an active knee orthosis using a computational model of the musculoskeletal system

Abstract

One-third of the stroke survivors remain with some disability, needing assistance to perform the activities of daily life and therapy to recover the lost functions. The robotic rehabilitation is a promissed field in this context improving the effectiveness of the treatment. Many researches have focused on developing human-robot interaction control to ensure user safety and therapy efficiency, but the validation of these controllers often requires contact between humans and robots, which involves cost, time and risk of accidents. This work aims to present a computational model of an ideal active orthosis used to assist the knee movement as a tool for test and validate human-robot interaction controls. Three controllers were applied to make the orthosis move the knee tracking the desired trajectory: a PID controller, an Inverse Dynamics-Based controller, and a Feedback-Feedforward Controller. The model proved to be useful and the controller with the best performance was the Feedback-Feedforward one.

A model-free deep reinforcement learning approach for control of exoskeleton gait patterns

“…Our mean absolute error was also lower than those reported in ref. [75], where a PID controller was applied to an orthotic attached to an OpenSim human musculoskeletal model with an average tracking error of 3.8 • . In comparison to other DRL-based exoskeleton control methods, which had mean absolute errors ranging from 0.12 • to 0.91 [86], respectively, our approach has comparable or better error (all our mean absolute errors were less than 0.67 • ), while uniquely allowing for model-free and accurate control for both seen and unseen gait patterns.…”

“…Passive forces from the muscle-tendon unit are incorporated into the simulation. To simulate a post-stroke individuals' baseline gait pattern, hip-knee-ankle joint torque patterns are applied to the joints of the musculoskeletal model through an additional torque actuator added to each musculoskeletal joint [75]. In following section, gait information data from post-stroke individuals are gathered to establish datasets of desired and baseline gait patterns.…”

The Effect of Lower Limb Exoskeleton Alignment on Knee Rehabilitation Efficacy

“…As an alternative to the risks associated with exposing humans to robot counterparts, computational models of an active orthosis have been utilized to assist knee movement during human–robot interaction (HRI) with promising results [ 10 ]. Musculoskeletal simulations could help identify the lower extremity functional range of motion which is accordingly used in the optimization of the synthesis procedure.…”