An investigation of the role of body-worn inertial sensing in the analysis of elite swimming performance

Abstract

Swimming is a technically demanding sport that requires ongoing quantitative assessment in order to monitor technical progression and improvements in performance. Traditional methods of monitoring a swimmer’s technique rely on the use of video-based systems. However, the primary motivation for this thesis is that these systems have several limitations when applied in aquatic environments. Such limitations are hindering the extent to which quantitative analytical practices are used by elite swimming coaches. As a consequence, alternative solutions are required and the advancement in the miniaturisation of microelectromechanical systems (MEMS) has led to a recent increase is the interest in applying such technology in swimming.

This thesis describes a set of studies focused on investigating the application of existing MEMS technology for the analysis of elite swimming performance. MEMS sensors such as accelerometers and gyroscopes have been shown to be capable of registering some basic parameters relevant to the analysis of swimming, such as lap time and stroke count, but further research and development are necessary in order to improve the functionality of these devices and to increase the applicability of this approach in elite settings.

This thesis also describes the development of a novel swimming analysis system, based on the use of MEMS inertial sensor technology. A user-centred design approach was followed to fully investigate current practices and to understand the challenges of incorporating this technology in applied training environments. A key contribution of this work is the development of a number of novel feature detection algorithms for the analysis of swimming turns. These studies demonstrate the feasibility of incorporating MEMS technology in an elite swimming environment to inform and enhance the coaching process.