On track

Technology used to help trace criminals may be used to boost UK athletes’ Olympic 2012 hopes


A technology being used to identify possible criminals or terrorists in a crowd may help UK in its bid for medals at the 2012 Olympics.

The analysis of body movement using artificial intelligence helps security services pick out unusual body movements and suspicious behaviour at events such as rallies or football matches.

Now, Muscle Memory Second Generation (MM2G) is developing a web-based product that will use the same technology to enhance the sports performance of professional and amateur athletes.

The system uses qualitative analysis based on technology developed with the Institute of Industrial Research (IIR) at the University of Portsmouth. The group has studied how the body’s muscles operate when the subject is performing a particular activity and, using artificial intelligence, has created computer templates of normal movement.

‘The theory of how this will work is that the sportsperson will capture their movements and compare it to the templates,’ said Winslie Gomez, MM2G managing director. ‘This will create the equivalent of having a personal coach in their pocket.’

By comparing normal movement to that of an individual, it should be possible to identify any potential problems that could compromise performance or even result in long-term injury.

Although the system is initially going to be aimed at the UK market, with the objective of helping athletes training for the 2012 Olympics in London, MM2G hopes eventually it will develop to become a global interactive e-learning and teaching tool.

‘The aim is to marry artificial intelligence and behavioural templates to sports science and performance. The resulting product will take sports training and performance analysis onto the internet,’ said Gomez.

‘We are going to be taking on a postgraduate student who will create the normalised templates of how the body behaves during sports, using artificial intelligence.’

MM2G will refine the IIR’s templates in order to use them in a sports context. The company is looking for a doctoral student to conduct research aimed at developing real-time and feasible algorithms for this recognition and matching of continuous human motion to the sports templates they will create.

Once the tool is ready, it is envisaged that one or more cameras will allow the user to capture images of their movements as they carry out activities. These will then be transmitted to a central computer using a device such as a PDA.

Part of the project will involve determining the minimum number of cameras that are needed to capture the data required. Such movement analysis will involve pulling several templates together. However, this will create a mass of information, as it will involve large amounts of image capture.

Even given the improvements in processing power that are expected over the next three years, in order to pick out the relevant portions and ensure these are still small enough to be sent from a personal device, an algorithm will need to be developed that can hone in on and filter out the data required.

As well as professional sports stars, Gomez said that players in, for example, amateur rugby teams could also benefit from its use. It could also have a role in creating an extensive online resource, similar to an encyclopaedia of sport that would contain interactive portions.

‘The level of analysis that the system would deliver could be tailored to the user’s needs,’ said Gomez.

The system also has the potential for use in treating and preventing industrial and work injuries, such as postural problems caused by workers slumping over keyboards for a long period.

‘We are walking into a disaster area of health when it comes to hunching over a computer,’ said Gomez. ‘By modifying movements we could trigger the muscles to change and improve themselves.’