A Novel Method for Detecting and Monitoring Multidirectional Running in Rugby League Referees

Abstract

<p>To date there has been limited research investigating the movement demands of rugby league referees and specifically the contribution of multidirectional movement to the overall locomotive demands. Previous methods utilised for detecting and quantifying multidirectional locomotion in rugby league referees have shown various limitations. Considering multidirectional locomotion has many biomechanical, physiological, and perceptual differences when compared to forward locomotion, valid methods are required for accurately understanding the overall movement demands placed on rugby league referees during match play. The first experimental chapter of this thesis aimed to address the limitations of previous research by utilising a novel approach for detecting multidirectional locomotion. Specifically, this chapter utilised key signal characteristics from microtechnology as inputs into a deep learning algorithm, with the results indicating that the algorithm was able to detect multidirectional locomotion during a simulated movement protocol in a group of rugby league referees. Following this, the aim of the second experimental chapter was to examine the feasibility of applying the novel algorithm derived in the first experimental study to match data of rugby league referees for detecting and quantifying multidirectional locomotion. In addition to this, it also provides a brief examination of the differences in movements demands between referee roles from a subset of data from a NRL season. The results of this chapter highlighted the practicality of the novel algorithm to elucidate multidirectional movement demands from match data of rugby league referees. Overall, the current thesis provides a novel approach to enhance load monitoring methods through utilising microtechnology and deep learning processes for detecting and quantifying multidirectional locomotion. </p>