Multidimensional ground reaction forces and moments from wearable sensor accelerations via deep learning

Abstract

Introduction: Military task performance is dependent upon the coherent and effective collaboration of its component parts. These are comprised of individuals working in coordinated teams performing tasks that are naturally variable and have physical and cognitive elements, the relative contribution being dependent upon the task. Furthermore, the battlefield is non-linear, and dynamic with human, task, environmental and informational constraints. Complexity is added to team performance assessment by the fact that there is considerable redundancy within military teams. This redundancy is critical and affords the ability for team performance to be protected even when individual performance is degraded by fatigue. The purpose of this work is to propose a conceptual framework for the assessment of military team dynamics, based upon analytical models used in sport in addition to emergent sensor technologies and Artificial Intelligence (AI). Methods: A review of the scientific literature and the developments in team performance assessment, AI and emerging sensor technologies was conducted to inform a conceptual framework designed to assess team dynamics and redundancy in the presence of individual fatigue. Results: Current assessment of team dynamics during mission-critical tasks, the contribution of each individual, and the emergent team behaviours whilst subject to fatigue are typically limited to subjective ratings and post-activity reports. Indeed, there seems to be an institutional belief that the military already assess teams effectively and retrospective After Action Review is key to this success. As such, the ability to reliably monitor and objectively assess military team dynamics and redundancy remains a challenge. In addition, despite advancements in technology, the opportunity to conduct real-time or near real-time observations with which to provide rapid insights is yet to be taken up by the military. Discussion/conclusions: Improvements in analytical methods, sensor technologies, and the maturity of AI tools provide the opportunity to progress from simply describing individual and team performance to understanding the team dynamics by the integration of novel sensors and AI-assisted analytical tools. Devices to rapidly acquire and process data continue to evolve along with wearable, and stand-off sensors and AI-assisted tools. A peer-reviewed, conceptual, interdisciplinary framework for objectively assessing team dynamics is proposed. With careful collaboration, we suggest that the potential to enhance the understanding of team dynamics and redundancy is boundless. Notwithstanding, it is critical that defence organisations have an endorsed strategy for the acceptance and implementation of such analytical tools that would provide a clear operational benefit.