Traditionally, in sports performance analysis, scientists often rely on expensive and bulky measuring equipment, such as optical based motion capture systems (similar to those used in animated films). Therefore, such analysis is mainly designed for elite athletes. Due to the inherent limitations of optical tracking technology (it requires a direct line of sight), the analysis can only be carried out in a laboratory environment and cannot be used outside of the laboratory. Following the evolution of the semiconductor industry, computers are getting smaller and smaller and, recently, a growing number of wearable computers have become available, such as the Apple iWatch and the Microsoft Band. Equipped with inertial and physiological sensors, these new wearable computers can enable detailed performance analysis in the field, and some wearable computers have already been used in professional sports, for example, in rugby teams for training and athlete performance analysis during competitions. The Hamlyn Centre at Imperial College jointly with Sensixa Limited has developed a range of wearable sensors for sport performance monitoring. Detailed performance parameters can be derived from these sensors to allow for fine tuning of techniques. The wireless sensors can transmit all the derived performance metrics to the coach or the user’s computer shortly after each training session and such information will help the coach and user to refine the training to target the areas that can lead to performance improvement. The low cost wearable sensor will not only enable in-field detailed performance analysis for elite athletes, but it can also be made affordable to anyone who has a keen interest in sports.

Swimming

Head movements captured by the sensor can be used to obtain detailed performance parameters, such as type of stroke, stroke rate, speed, etc. for swimming. The major challenges in using pervasive sensing technologies for swimming performance analysis are to minimise the obtrusion to the swimmer, the limitation of the wireless range, the loss of wireless signal underwater, and the relevance of the performance metric derived from the wearable sensor. To limit the obtrusion to the athlete, the miniaturised waterproof ear worn sensor can be worn underneath the swimming cap, and since it is based on an inertial sensor, it can also be integrated into a swimming goggle. However, the miniaturised sensor has limited computational power and limited wireless range. Based on 2.4GHz, most of the radio signal will be lost underwater. To overcome the limitation on the radio range, a low power opportunistic asynchronous communication protocol (OADC) was developed by the research team. By utilising the on-node memory, the OADC can store the data in the memory and transmit it to a computer via a base station wirelessly when the sensor is in range. In addition, a visual performance metric was designed to show the individual performance in a spider plot highlighting the stroke rate, stroke length and speed. The swimmer can easily spot the improvements, weakness etc., compare to previous swims or to other elite athletes right after each swim, and targeted improvements can be made subsequently.  

Rowing

In professional rowing, quantitative analysis is mainly carried out in laboratory settings by using ergometers, as it is difficult to instrument a rowing boat with traditional biomechanical and physiological measurement tools. To facilitate the capturing and visualising real-time information from body worn sensors to boat sensors, an ESPRIT Blackbox is designed and developed.  By integrating the inertial sensors, GPS and low power wireless interface, the ESPRIT Blackbox enables real-time capturing of rowing performance indices, such as boat speed, acceleration, stroke rate, etc. In addition, it can also link wirelessly to the on-body sensors on the rowers to capture the physiological and bio-motion indices. With the built-in mobile network connectivity, the ESPRIT Blackbox provides the necessary bridge for the body worn sensors and on boat sensors to the data server , thus enabling real-time visualisation and performance profiling.