Optimising load carriage performance and reducing injury risk in military personnel across undulating terrain.

Load carriage is a routinely performed, demanding military task [1]. However, this task imposes biomechanical, physiological, and mental stress that can hinder performance and increase the risk of injury [2–4]. Measuring task demands requires the use of field-expedient wearable sensors, such as inertial measurement units (IMUs) and wrist-worn devices (e.g., Garmin watches). Employing sensors during training can identify key factors associated with optimal load carriage performance and injury using machine learning (ML) models.

AIMS
Identify biomechanical, physiological, and resilience factors associated with load carriage performance and injury risk using data from a 6.72 km time trial and ML models.
Cross-validate the developed ML models to evaluate whether predictors from Aim 1 generalise to unloaded training marches.
SIGNIFICANCE

Although evidence shows that load carriage in laboratory settings is associated with higher ground reaction forces, forward trunk lean/centre-of-mass shift, gait variability, and increased heart rates compared to unloaded conditions, it is unclear which biomechanical, physiological, and resilience factors are associated with outdoor load carriage performance and potential injury risk when using ML techniques. Field-related conditions, including gradients, terrain, and task intensity, can impact task performance. Together, these conditions remain underinvestigated and are often not considered in ML models when evaluating the role of biomechanical, physiological, and resilience metrics in military performance and injury risk [3, 5–8]. Employing field-based sensors is essential for determining optimal load carriage task characteristics, providing greater ecological validity. These data may assist commanders in making strategic, real-time decisions to adjust training prescriptions, identify personnel exhibiting at-risk signatures, and optimise mission success.

OUTCOMES
A preliminary ML model that discerns the role of wearable-derived biomechanical, physiological, and resilience characteristics in predicting performance outcomes and injury risk during a time-trial load carriage task.
Predictors identified during the loaded task (Aim 1) will be reliable predictors of performance and injury risk during unloaded marches.

Funder: UNE Internal Funding Scheme for Early Career Researchers (ECRs) and Mid-Career Researchers (MCRs)
Awarded amount: A$5,000.00