Novel Methods for Personalized Gait Assistance: Three-Dimensional Trajectory Prediction Based on Regression and LSTM Models.

Enhancing human-robot interaction has been a primary focus in robotic gait assistance, with a thorough understanding of human motion being crucial for personalizing gait assistance. Traditional gait trajectory references from Clinical Gait Analysis (CGA) face limitations due to their inability to account for individual variability. Recent advancements in gait pattern generators, integrating regression models and Artificial Neural Network (ANN) techniques, have aimed at providing more personalized and dynamically adaptable solutions. This article introduces a novel approach that expands regression and ANN applications beyond mere angular estimations to include three-dimensional spatial predictions. Unlike previous methods, our approach provides comprehensive spatial trajectories for hip, knee and ankle tailored to individual kinematics, significantly enhancing end-effector rehabilitation robotic devices. Our models achieve state-of-the-art accuracy: overall RMSE of 13.40 mm and a correlation coefficient of 0.92 for the regression model, and RMSE of 12.57 mm and a correlation of 0.99 for the Long Short-Term Memory (LSTM) model. These advancements underscore the potential of these models to offer more personalized gait trajectory assistance, improving human-robot interactions.

Individualized Three-Dimensional Gait Pattern Generator for Lower Limbs Rehabilitation Robots.

In the field of robotic gait rehabilitation, controlling robotic devices to follow specific human-like trajectories is often required. In recent years, various gait generator models have been proposed, providing customized gait patterns adjustable to a range of heights and gait speeds. However, these models were developed with a focus on gait rehabilitation devices designed to control the angular trajectories of the subject's joints, e.g. exoskeletons. Similar devices, e.g. end-effector robots, control the orientation and also the 3D position of the subject's joints and cannot easily implement these models. In this study, it is proposed a new individualized three-dimensional gait pattern generator for gait rehabilitation robots. The generator employs multi-variable regression models to predict the joint angular trajectories of the pelvis, hip, and ankle along the gait cycle. The 3D joints positions are then reconstructed by applying the predicted angular trajectories over a human model inspired on the inverted pendulum analogy using inverse kinematics. The generator's performance was statistically evaluated against real gait patterns from 42 participants walking at 8 different velocities. The predicted trajectories matched the measured ones with an average Root Mean Squared Error of 25.73 mm for all joints at all Cartesian axes, with better results between 3.3 - 5.4 km/h. Suggesting to be a good solution to be applied in end-effector gait robotic rehabilitation devices.