Gait Trajectory and Gait Phase Prediction Based on an LSTM Network.

Binbin Su,Elena M Gutierrez-Farewik

doi:10.3390/s20247127

Binbin Su, Elena M Gutierrez-Farewik

Open Access

https://doi.org/10.3390/s20247127

Copy DOI

Abstract

Lower body segment trajectory and gait phase prediction is crucial for the control of assistance-as-needed robotic devices, such as exoskeletons. In order for a powered exoskeleton with phase-based control to determine and provide proper assistance to the wearer during gait, we propose an approach to predict segment trajectories up to 200 ms ahead (angular velocity of the thigh, shank and foot segments) and five gait phases (loading response, mid-stance, terminal stance, preswing and swing), based on collected data from inertial measurement units placed on the thighs, shanks, and feet. The approach we propose is a long-short term memory (LSTM)-based network, a modified version of recurrent neural networks, which can learn order dependence in sequence prediction problems. The algorithm proposed has a weighted discount loss function that places more weight in predicting the next three to five time frames but also contributes to an overall prediction performance for up to 10 time frames. The LSTM model was designed to learn lower limb segment trajectories using training samples and was tested for generalization across participants. All predicted trajectories were strongly correlated with the measured trajectories, with correlation coefficients greater than 0.98. The proposed LSTM approach can also accurately predict the five gait phases, particularly swing phase with 95% accuracy in inter-subject implementation. The ability of the LSTM network to predict future gait trajectories and gait phases can be applied in designing exoskeleton controllers that can better compensate for system delays to smooth the transition between gait phases.

Highlights

Lower limb exoskeletons (LLEs) have for decades received increasing attention and interest in the field of rehabilitation for patients with a disability affecting their ability to locomote bipedally
The objectives of this study are to develop an approach for multiple-frame prediction of gait trajectory and gait phases based on a long short-term memory (LSTM) network in which a discount factor in the loss function is introduced
They were equipped on each foot with 4 foot switches, small piezoresistive pressure sensors taped to the soles of the feet at the heels, first and fifth metatarsal heads and big toe, that record an on or off signal based on ground contact

Summary

Introduction

Lower limb exoskeletons (LLEs) have for decades received increasing attention and interest in the field of rehabilitation for patients with a disability affecting their ability to locomote bipedally. They can provide vertical support, assistance in lower limb motion, and propulsion of the body to help wearers regain locomotion function. If a gait trajectory can be predicted and incorporated into the control algorithm, it may be beneficial to add a feed-forward component to compensate for the delay that may result in the control’s response time [2,3,4]. Both model-based and machine learning methods are widely studied approaches for gait trajectory prediction

Objectives

Methods

Results

Discussion

Conclusion