Professor uses mobiles to ‘predict the future’

23 Aug 2005

A Sligo-based computing lecturer has devised a way of predicting where a person is going to be in the future based on the pattern of a person’s mobile phone use.

When tested against records on mobile users’ actual movements, Dr Paul Flynn’s model’s predictions were found to be accurate 93pc of the time.

As well as helping mobile networks pass on their customer’s messages, the model could also help shopkeepers predict when a person is likely to walk past their window – giving them enough time to put together an advertisement or display targeted at that person.

Flynn analysed some four million mobile call records to test the accuracy of his prediction model. They showed its predictions on the whereabouts of mobile users were 93pc accurate, a far higher accuracy level than many other models.

Flynn, who lectures in computing at Institute of Technology Sligo, was recently conferred with a PhD by the University of Ulster for his six-year research project.

At present, networks have to find out where a mobile phone user is in order to deliver a message. Current methods such as broadcasting the information through paging or using a location update signal from the mobile phone eats up too much bandwidth, power and data storage. Researchers are looking for ways to cut down on the amount of signalling traffic that’s needed.

The next-generation mobile technology will have a single wireless handset that will provide several different services using high-speed wireless networks. There will be a convergence of fixed-line voice and data and mobile voice and data into one, handheld, wireless device.

The two most important issues for the new high-speed networks will be user management and location prediction. To predict a person’s location involves a number of strategies. One element of this is to use the mobile user’s pattern of past use and location to predict where they are likely to be at a given time. If this problem could be solved it would slash the amount of signalling traffic needed to pass on calls.

Flynn’s research involved the examination of all contributory factors that could be used to accurately predict a future location of a mobile wireless user in a future micro-mobile environment. He designed a multifaceted model that took all elements of mobile movement into consideration.

The Resource Allocation Mobility Location Prediction Model uses mathematical formulae to analyse past movement history and special artificial intelligent network software, as well as other prediction information, such as random user movements, channel allocation, traffic profiles and signal deployment, to make the most accurate prediction of a mobile user’s whereabouts.

Flynn hopes to continue research in this field, particularly on micro and pico-cellular mobile environments.

By John Kennedy