Saving lives, bit by bit

Developing new methods for distributed intelligent systems to share and analyse digital information was a huge academic challenge –
and its solution might save lives.

ALADDIN, or Autonomous Learning Agents for Decentralised Data and Information Networks,  was an ambitious five-year project which brought together leaders in information processing to work closely with industry. Oxford’s contribution came from Professor Stephen Roberts’ Pattern Analysis and Machine Learning Research Group who, alongside BAE Systems and researchers from other UK universities, merged strands of information engineering that had never been combined before.

One major focus of the project was disaster management, a real-world problem where there is an urgent need to co-ordinate activities in the face of great uncertainty: from allocating emergency services to finding the best escape routes. The challenge was how to take swathes of data available in these situations and use it in some way on the ground, in real time, to provide the best outcome.

To do that required a combination of game theory, probabilistic modelling and optimisation techniques to make the most effective use of all the available data. The team from Oxford took a crucial role in the project, working on how to combine and share around the network the data from vast numbers of sensors – a process they refer to as data fusion, and one that is used across the entire project.

The project saw the researchers develop intelligent ways  to form and deploy the most effective rescue teams and dynamic systems for evacuating buildings that always provide the safest escape option, amongst others. They also developed ways to overcome problems of faulty sensors, system delays and disrupted communication, while still achieving the desired results.

The collaboration ran from 2005–10, producing more than 150 publications, winning The Engineer’s Award for Aerospace and Defence in 2009, and generating three joint patents held between BAE systems and the University of Oxford.

The research has even spawned another five-year research program, called ORCHID. The project, funded by EPSRC and BAE Systems with over £10 million, aims to utilise humans as extra ‘computational cores’, by providing them with appropriate and  timely information,  to improve the human contribution to areas as diverse as disaster management, energy conservation and crowd-sourced citizen science.

Funded by: The Engineering and Physical Sciences Research Council, BAE Systems.