Predicting ground reaction forces in running using micro-sensors and neural networks

Billing, Daniel C.; Nagarajah, Romesh; Hayes, Jason P.; Baker, John

doi:10.1007/bf02844259

Cited by 13 publications

(10 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An ankle joint torque MR prediction model has already been developed with adequate accuracy (R 2 ADJUSTED = 0.831, RMSE = 6.91 Nm) using the independent input of 99-sensor pressure insoles [ 54 ]. Further, vertical GRF from pressure insoles have been used to predict the 3D GRF components using MR and Artificial Neural Networks, supporting that power and injury related variables can be considered a possibility via simple wearable sensors [ 55 ]. From an application standpoint, although many independent variables are used in the models of the current study, they all are derived from a single system.…”

Section: Discussionmentioning

confidence: 99%

Foot Strike Angle Prediction and Pattern Classification Using LoadsolTM Wearable Sensors: A Comparison of Machine Learning Techniques

Moore

Kranzinger

Fritz

et al. 2020

Sensors

View full text Add to dashboard Cite

The foot strike pattern performed during running is an important variable for runners, performance practitioners, and industry specialists. Versatile, wearable sensors may provide foot strike information while encouraging the collection of diverse information during ecological running. The purpose of the current study was to predict foot strike angle and classify foot strike pattern from LoadsolTM wearable pressure insoles using three machine learning techniques (multiple linear regression―MR, conditional inference tree―TREE, and random forest―FRST). Model performance was assessed using three-dimensional kinematics as a ground-truth measure. The prediction-model accuracy was similar for the regression, inference tree, and random forest models (RMSE: MR = 5.16°, TREE = 4.85°, FRST = 3.65°; MAPE: MR = 0.32°, TREE = 0.45°, FRST = 0.33°), though the regression and random forest models boasted lower maximum precision (13.75° and 14.3°, respectively) than the inference tree (19.02°). The classification performance was above 90% for all models (MR = 90.4%, TREE = 93.9%, and FRST = 94.1%). There was an increased tendency to misclassify mid foot strike patterns in all models, which may be improved with the inclusion of more mid foot steps during model training. Ultimately, wearable pressure insoles in combination with simple machine learning techniques can be used to predict and classify a runner’s foot strike with sufficient accuracy.

show abstract

Section: Discussionmentioning

confidence: 99%

Foot Strike Angle Prediction and Pattern Classification Using LoadsolTM Wearable Sensors: A Comparison of Machine Learning Techniques

Moore

Kranzinger

Fritz

et al. 2020

Sensors

View full text Add to dashboard Cite

show abstract

“…The in-shoe load sensors were commercially available (paromed Vertriebs GmbH & Co. KG [9]) MEMS piezoelectric micro-sensors embedded into silicone-filled hydrocells [5]. The discrete in-shoe sensors were deployed at a site approximating the heel, first metatarsal head (MTH), third MTH and hallux of both the left and right foot.…”

Section: Equipmentmentioning

confidence: 99%

“…They then combined the data from the different sensors via neural networks to predict the three dimensional ground reaction forces [5].…”

Section: Introductionmentioning

confidence: 99%

Validation of trunk mounted inertial sensors for analysing running biomechanics under field conditions, using synchronously collected foot contact data

2010

Self Cite

View full text Add to dashboard Cite

The biomechanical evaluation of elite athletes often requires the use of sophisticated laboratory-based equipment that is restrictive, cumbersome, and often unsuitable for use in a training and competition environment. Small, low-mass unobtrusive centre-of-mass triaxial accelerometers can be used to collect data but may not reveal all the information of interest. This validation of centre-of-mass triaxial accelerometry uses previously reported synchronously collected foot-contact information from in-shoe pressure sensors. A qualitative assessment of the system output indicates that the centre-of-mass acceleration provides valuable insight into the use of accelerometers for investigating the biomechanics of, in this case, middle distance runners.

show abstract

“…However, it is clear that many measurement techniques are not routinely available, nor routinely used -and the reasons for this have not been investigated but are considered to include cost, locality, knowledge and expertise of the coach (and athlete) and general support for these activities by the sport governing body/funding body. It may also be that whilst many of the more advanced techniques have been used at research level (Billing et al, 2006), there is still further work required to assess the accuracy and repeatability of measurements (Putti et al, 2007), and the way they record and present data compared to the requirements of the user. As a very simple example of poor accuracy the Nike pod (fitted to a user's shoe) is initially calibrated to the user's typical stride length during a single short run and then assumes that all running strides thereafter are the same length as the calibrated stride in further measurements and analysis.…”

Section: Discussionmentioning

confidence: 99%

Athlete and coach perceptions of technology needs for evaluating running performance

Fleming

Young²,

Dixon

et al. 2010

Sports Eng

View full text Add to dashboard Cite

Athletes and their support team utilise technology to measure and evaluate technique and athletic performance. Existing techniques for motion and propulsion measurement and analysis include a combination of indirect methods (high-speed video) and direct methods (force plates and pressure systems). These methods are predominantly limited to controlled laboratory environments (in a small area relative to the competition environment), require expert advice and support, and can take significant time to evaluate the data.Consequently, the more advanced measurement techniques are considered to be restricted to specific coaching sessions, or periods in the year leading up to competition, when the time and expertise of further support staff are available. The more widely used, and simple, devices for monitoring 'performance' during running include stopwatches, GPS tracking and accelerometer-based systems to count strides. These provide useful information on running duration, distance and velocity but lack detailed information on many key aspects of running technique. In order to begin the process of development of more innovative technologies for routine use by athletes and coaches, a study was required to improve the understanding of athletes' and coaches' perception of their requirements from measurement technology. This study outlines a systematic approach to elicit and evaluate their perceptions, and presents the findings from interviews and a questionnaire. The qualitative data are presented as a hierarchical graphical plot (structured relationship model) showing six general dimensions (technique, footwear and surface, environment, performance, injury and cardiovascular) and shows the development of these general dimensions from 2 the interviewee quotations. The questionnaire quantitative data enhances the study by further ranking characteristics that arise from the interviews. A contrast is shown between short and longer distance runner groups, as might be expected. The current technology available to elite runners is briefly reviewed in relation to the 22 characteristics identified as important to measure. The conclusions highlight the need for newer technologies to measure aspects of running style and performance in a portable and integrated manner, with suggestions as to size and weight likely to be acceptable to users for emerging devices.

show abstract

Predicting ground reaction forces in running using micro-sensors and neural networks

Cited by 13 publications

References 11 publications

Foot Strike Angle Prediction and Pattern Classification Using LoadsolTM Wearable Sensors: A Comparison of Machine Learning Techniques

Foot Strike Angle Prediction and Pattern Classification Using LoadsolTM Wearable Sensors: A Comparison of Machine Learning Techniques

Validation of trunk mounted inertial sensors for analysing running biomechanics under field conditions, using synchronously collected foot contact data

Athlete and coach perceptions of technology needs for evaluating running performance

Contact Info

Product

Resources

About