With the rapid advances in sports technologies, athletes and sports coaches are constantly searching for improved performance assessment methods. Whilst athletic performances continue to improve, accurate training prescription and feedback is important to the consistency of the training outcome and maintaining the performance margin. To maximise the potential of UK athletes at future Olympics, Olympic Winter Games, and Paralympics, there is a pressing need to exploit the latest technical advances in sensing, materials, aerodynamics, biomechanics, and performance equipment design. In supporting the quest for gold in the London Olympics and Paralympics in 2012, UK Sports and EPSRC have identified a range of engineering and physical sciences disciplines that through the interaction with the sports community can generate innovative training solutions and sports equipment designs for gaining competitive advantage of the UK athletes. Such a synergy brings the opportunity not just to ensure success at the Games themselves, but to provide a sporting legacy that will underpin the long-term health and success of sport in this country. The purpose of this proposal is to investigate the use of miniaturised wireless Body Sensor Networks (BSN) for providing real-time feedback and in situ analysis of the biomechanical indices of the athletes during training. It is a feasibility project aimed at addressing the technical requirement of Sports-BSN hardware design, miniaturisation, packaging, as well as real-time data processing, sensor fusion, and data visualisation issues. The project brings together an interdisciplinary team from the Department of Computing at Imperial College London, UK Sport and nominated technical expertise working within the elite sport network (Dr Aki Salo, UK Athletics Speed specialist currently based at University of Bath).

A major problem in many physiological measurements is that the act of performing the measurement can itself alter the system that is under observation. The challenge for scientists is to develop techniques which allow the non invasive assessment of physiological processes. A technique called Near Infrared Spectroscopy (NIRS) has previously been used to measure the wavelength dependence of the optical absorption of blood and therefore its oxygen content (highly oxygenated arterial blood is bright red whilst oxygen depleted venous blood appears purple/blue in colour). In principle NIRS allows for the measurement of the oxygen saturation of the muscle. Muscles use oxygen to assist in the conversion of food energy (carbohydrate/fat) into the useable chemical energy that can drive muscle contraction and allow an athlete to run, cycle and swim. Exercise uses up oxygen and therefore how much oxygen is in the muscle (its oxygen saturation) is a measure of whether the oxygen being delivered is keeping up with its consumption. In aerobic exercise there is sufficient oxygen; in anaerobic exercise this is not the case. Achieving gold in the 2012 Olympics has become a top priority for UK Sport. There are many factors that will influence the position of Team GB in the medals table in 2012, including the development of optimised training regimes for elite athletes. Informed development of effective training strategies requires coaches to be given real time feedback on an athlete's performance at the trackside. Currently there is a scarcity of available devices that provide reliable and accurate physiological monitoring of elite athletes in the field. There are currently very few available methods which allow us to measure physiological changes in an athlete while they are training in the field. In theory NIRS methods could be used to measure muscle oxygenation in training athletes. However current commercial NIRS machines are large, heavy and non-portable.The aim of this feasibility project is to develop a non-obtrusive, battery driven, compact, NIRS device that measures local absolute muscle oxygen saturation and transmits this data via a wireless link to the coach in real time. Feedback from elite athletes and coaches at Essex (and via UK sport) will inform the design to ensure that the device will be acceptable to all athletes and will not compromise athletic performance. In parallel with the design of the prototype device we will be testing how the new technology could be used to enhance sports performance. We will focus initially on events, such as running and cycling, where it is easy to make complex biochemical and physiological measurements in the laboratory (using analysis of the exhaled air and measurements in blood samples). Our aim is to understand in detail the use of NIRS measurements in this laboratory setting using tests that attempt to recreate sporting events. Armed with this knowledge we will ultimately be able to apply the technology in the field during the training of an elite athlete. Of many possibilities we will initially focus on two uses for NIRS / optimising warm-up for athletes in sprint events and designing optimal pacing strategies (when to go fast and when to conserve energy) in endurance events. Beyond this feasibility study our long-term aim is to develop a device capable of providing a number of physiological measures for use in a wide range of sports and extreme environments including swimming and altitude training. Whilst we feel this device has the possibility of giving UK sport an edge in 2012, its manufacture will also be part of the Olympic post-2012 dividend, as it will have significant use in medical applications e.g. assisting paraplegics in improving muscle function and in investigating brain injury in patients.

To improve evaluation of physical performance and stamina in athletes undergoing rigorous training, it is necessary to expand our current armoury of physical monitoring tools beyond say pulse rate and respiratory gas exchange. We also lack true biochemical measures of physical endurance at the trackside; the problem has been partly that tissue and blood access are not acceptable in a well subject and also that simplified assay is not available. This proposal utilises thin film fabrication tools create multilayer electrode strips able to respond to O2 and to lactate. The former will be based on O2 going through intact skin by an in situ heater augment O2 transport out from the skin surface, and the latter will respond to lactate in saliva by means of an enzyme to break down lactate to generate a detectable hydrogen peroxide. There is accumulating data that salivary lactate, though low, reflects blood values. The overall aim is to achieve non-invasive, reagentless devices for sports medicine use. Specifically, transcutaneous O2 will give a measure of peripheral tissue O2 delivery and lactate will give a measure of the level of the tissue oxygen deficit during extreme activity.

The cumulative effect of training has been calculated to be the single most important factor influencing subsequent cycling performance. Given its significance, it is surprising that relatively little research attention has been directed to optimising this process. Previous attempts have been made to model training mathematically. These models have used some simple assumptions about the effect of training on performance and have been moderately successful, but only in the context of the laboratory. However, these studies have not succeeded in assisting coaches working with elite athletes as the mathematical models are too simple. As a consequence, the current practice of training cyclists relies simply upon coaches' intuition and experience. Recently it has become possible to measure the work done by cyclists directly from their bicycles. Consequently there is now a large database of detailed data on the work performed by racing cyclists in training and competition. This large amount of detailed, and accurate information on the amount of work that riders have completed in training and racing is unique to cycling and not possible to measure in athletes involved in other sports. Therefore, this study will take a new approach to modelling training by attempting to use this available training to try and learn more about the training process. The successful findings of this research will benefit British Cycling's riders and coaches, in the form of improved performance and more medals won. The findings will also help reduce under-achievement, and provide a greater understanding of the nature of training and the athletes' response. Finally, the results may also be of interest to all people looking to make the most of the beneficial effects of exercise.

There is now widespread recognition that it is possible to extract previously unknown knowledge from datasets using machine learning techniques. In particular, rule induction algorithms capture the structure of data in a form directly amenable to human understanding. This project will explore the utility of a form of rule induction which combines evolutionary computation with reinforcement learning to produce human-readable solutions to facilitate knowledge discovery with respect to race analyses of the British swimming team. The technique, known as the Learning Classifier System (LCS), has recently shown great potential for data mining problems. LCS, along with other machine learning approaches, will be used to explore athlete datasets provided by the collaborating coach of the British swimming team.