Continuous Classification of Locomotion in Response to Task Complexity and Anticipatory State.

Mahdieh Kazemimoghadam,Nicholas P Fey

doi:10.3389/fbioe.2021.628050

Abstract

ObjectiveIntent recognition in lower-extremity assistive devices (e.g., prostheses and exoskeletons) is typically limited to either recognition of steady-state locomotion or changes of terrain (e.g., level ground to stair) occurring in a straight-line path and under anticipated condition. Stability is highly affected during non-steady changes of direction such as cuts especially when they are unanticipated, posing high risk of fall-related injuries. Here, we studied the influence of changes of direction and user anticipation on task recognition, and accordingly introduced classification schemes accommodating such effects.MethodsA linear discriminant analysis (LDA) classifier continuously classified straight-line walking, sidestep/crossover cuts (single transitions), and cuts-to-stair locomotion (mixed transitions) performed under varied task anticipatory conditions. Training paradigms with varying levels of anticipated/unanticipated exposures and analysis windows of size 100–600 ms were examined.ResultsMore accurate classification of anticipated relative to unanticipated tasks was observed. Including bouts of target task in the training data was necessary to improve generalization to unanticipated locomotion. Only up to two bouts of target task were sufficient to reduce errors to <20% in unanticipated mixed transitions, whereas, in single transitions and straight walking, substantial unanticipated information (i.e., five bouts) was necessary to achieve similar outcomes. Window size modifications did not have a significant influence on classification performance.ConclusionAdjusting the training paradigm helps to achieve classification schemes capable of adapting to changes of direction and task anticipatory state.SignificanceThe findings could provide insight into developing classification schemes that can adapt to changes of direction and user anticipation. They could inform intent recognition strategies for controlling lower-limb assistive to robustly handle “unknown” circumstances, and thus deliver increased level of reliability and safety.

Highlights

Locomotion is initiated by robust motor patterns within the nervous system and involves synchronized activity of muscles and the skeletal system triggering a set of cyclic movements (Chiel and Beer, 1997)
Higher levels of accuracy were reported for anticipated mixed transitions (A-crossover to stair-ascent (COS), A-sidestep to stair-ascent (SSS)) and straight walking compared to single transitions (A-CO, A-SS) which was mainly due to misclassification of A-CO and A-SS as each other (∼13%) (Table 1)
A-SS and A-SSS error rates decreased from 35% at the beginning of the trial to

Summary

Introduction

Locomotion is initiated by robust motor patterns within the nervous system and involves synchronized activity of muscles and the skeletal system triggering a set of cyclic movements (Chiel and Beer, 1997). Advanced lower-limb assistive devices (i.e., prostheses and exoskeletons) are being developed to better aid individuals with motor impairments/amputation in performing daily activities. Microcontroller-based/powered assistive devices have the potential to provide intuitive transitions between locomotor tasks and allow automatic and smooth “steering” (Huang et al, 2011; Young and Ferris, 2017). In order to achieve seamless integration of the device with the wearer, gait information needs to be acquired real time, and locomotor mode should be instantaneously identified by the control algorithms (Varol et al, 2010; Jiménez-Fabián and Verlinden, 2012). Intent recognition using machine learning has been the popular methodology to infer user intention for the device, and to effectively identify target locomotion modes (Young et al, 2013; Hargrove et al, 2015)

Methods

Results

Discussion

Conclusion