A visualisation of functional principal components analysis

A series of four graphs


With the advent of high-fidelity sensors and imaging systems, the applied sports sciences have observed a huge influx of available data originating from different sources, including wearable sensors that are now ubiquitous across many sporting disciplines. These sensors can quantify dynamic human movements over extended periods of time to enable reliable monitoring of athletic performance.

Development of system models without supporting sensor measurements in sporting data is inherently difficult; even the simplest of sporting situations can give rise to non-linear behaviour. For example, the dynamics of a limb, an individual or team members can abruptly change in anticipation of, or response to, a perceived scoring opportunity or an opponent’s action. Consequently, first principle models of most sporting systems are challenging to formalise. However, the confluence of high-fidelity measurements (Big-Data) and the advancement of numerical methods provides the means to account for the non-linearity, high-dimensionality, multiscale nature, and intrinsic uncertainty observed in most sporting systems.

Our specialist area aims to take advantage of Big-Data and applied mathematical and statistical techniques to discover underlying sporting system models. We employ techniques such as regression, machine learning, neural networks and genetic programming to formalise unknown dynamics and provide a means to predict the future state of the sporting system in question. Additionally, we strive to ensure that data-driven models are interpretable and intuitive so the discovered system can be physically understood by practitioners working in applied sporting environments.


Specialisation: Biomechanics, measurement and sensors
Specialisation: Complex dynamical systems, signal processing
Specialisation: Performance science
Specialisation: Dynamical systems, control theory
Specialisation: Physiology, human biology, sports medicine

Current Projects