Sprint Assessment Using Machine Learning and a Wearable Accelerometer

Machine learning methods have been widely used for gait assessment through the estimation of spatio-temporal parameters. As a further step, the objective of this work is to propose and validate a general probabilistic modeling approach for the classification of different pathological gaits. Specifically, the presented methodology was tested on gait data recorded on two pathological populations (Huntington’s disease and post-stroke subjects) and healthy elderly controls using data from inertial measurement units placed at shank and waist. By extracting features from group-specific Hidden Markov Models (HMMs) and signal information in time and frequency domain, a Support Vector Machines classifier (SVM) was designed and validated. The 90.5% of subjects was assigned to the right group after leave-one-subject–out cross validation and majority voting. The long-term goal we point to is the gait assessment in everyday life to early detect gait alterations.

The purpose of this study was to assess validity and reliability of sprint performance outcomes measured with an iPhone application (named: MySprint) and existing field methods (i.e. timing photocells and radar gun). To do this, 12 highly trained male sprinters performed 6 maximal 40-m sprints during a single session which were simultaneously timed using 7 pairs of timing photocells, a radar gun and a newly developed iPhone app based on high-speed video recording. Several split times as well as mechanical outputs computed from the model proposed by Samozino et al. [(2015). A simple method for measuring power, force, velocity properties, and mechanical effectiveness in sprint running. Scandinavian Journal of Medicine & Science in Sports. https://doi.org/10.1111/sms.12490] were then measured by each system, and values were compared for validity and reliability purposes. First, there was an almost perfect correlation between the values of time for each split of the 40-m sprint measured with MySprint and the timing photocells (r = 0.989-0.999, standard error of estimate = 0.007-0.015 s, intraclass correlation coefficient (ICC) = 1.0). Second, almost perfect associations were observed for the maximal theoretical horizontal force (F0), the maximal theoretical velocity (V0), the maximal power (Pmax) and the mechanical effectiveness (DRF - decrease in the ratio of force over acceleration) measured with the app and the radar gun (r = 0.974-0.999, ICC = 0.987-1.00). Finally, when analysing the performance outputs of the six different sprints of each athlete, almost identical levels of reliability were observed as revealed by the coefficient of variation (MySprint: CV = 0.027-0.14%; reference systems: CV = 0.028-0.11%). Results on the present study showed that sprint performance can be evaluated in a valid and reliable way using a novel iPhone app.

Gait speed is a powerful clinical marker for mobility impairment in patients suffering from neurological disorders. However, assessment of gait speed in coordination with delivery of comprehensive care is usually constrained to clinical environments and is often limited due to mounting demands on the availability of trained clinical staff. These limitations in assessment design could give rise to poor ecological validity and limited ability to tailor interventions to individual patients. Recent advances in wearable sensor technologies have fostered the development of new methods for monitoring parameters that characterize mobility impairment, such as gait speed, outside the clinic, and therefore address many of the limitations associated with clinical assessments. However, these methods are often validated using normal gait patterns; and extending their utility to subjects with gait impairments continues to be a challenge. In this paper, we present a machine learning method for estimating gait speed using a configurable array of skin-mounted, conformal accelerometers. We establish the accuracy of this technique on treadmill walking data from subjects with normal gait patterns and subjects with multiple sclerosis-induced gait impairments. For subjects with normal gait, the best performing model systematically overestimates speed by only 0.01 m/s, detects changes in speed to within less than 1%, and achieves a root-mean-square-error of 0.12 m/s. Extending these models trained on normal gait to subjects with gait impairments yields only minor changes in model performance. For example, for subjects with gait impairments, the best performing model systematically overestimates speed by 0.01 m/s, quantifies changes in speed to within 1%, and achieves a root-mean-square-error of 0.14 m/s. Additional analyses demonstrate that there is no correlation between gait speed estimation error and impairment severity, and that the estimated speeds maintain the clinical significance of ground truth speed in this population. These results support the use of wearable accelerometer arrays for estimating walking speed in normal subjects and their extension to MS patient cohorts with gait impairment.