The ALADDIN project aims to improve the way that emergency services respond in the event of a disaster
category/aerospace & defence
winner/aladdin/southampton university/oxford university/imperial college/bae systems
Co-ordinating disaster response in a city is a dizzyingly complex operation. There’s no time to think, but thinking is imperative, to make sure that the right emergency services get to the right place at the right time and respond as the situation develops. ALADDIN – Autonomous Learning Agents for Decentralised Data and Information Networks – is a multi-centre, multi-disciplinary consortium that will set out to find solutions to these, as well as related, problems.
Central to ALADDIN is the development and manipulation of ‘agents’ – systems combining sensors, software and communications components that play some part in the management of a larger system. Police cars can be agents, as can any other emergency vehicle; the sensors on buildings can be agents; the emergency services themselves can be agents. In a large system, various agents can have completely different owners and operators, and the uncertainty between data from different agents can vary considerably, but nonetheless, all the agents have to work together.
ALADDIN grew out of a previous project called Argus 2, which looked at situations such as the monitoring of airport and airfield tarmac, taxiways and runways. ‘The core bits of Argus 2 were Oxford and Southampton and BAE Systems,’ said project director Nick Jennings of the Southampton School of Electronics and Computer Science. ‘I was extremely proud of what we achieved there, but there was more to do; we want to explore the promising connections between dealing with uncertainty – using logical Bayesian reasoning, which Oxford deals with – and the multi-agent systems that we work on,’ he said.
Moving forward to ALADDIN, the teams decided to work on disaster management as the main focus of the project. ‘It was initially inspired by working out how emergency services could best respond to earthquakes, but it could also be applied to other situations, particularly floods and terrorist incidents: something that happens on a large scale and develops across the area of the city,’ added Jennings. ‘We’re looking at how the responders to the incident – the police, the fire brigades and ambulances – can work out how to best respond to what’s going on.’
The way the situation develops and the capabilities of each agent are key parts of the system. Some fire engines will be capable of handling certain sorts of fire better than others, Jennings pointed out, and the agents have to co-ordinate with each other. ‘If you have a burning building, you need appropriately equipped fire engines to deal with the fire and get people out; you need ambulances to treat the wounded and you need police to cordon areas off and control the access. As fires spread through the city, you need to re-co-ordinate on the fly.’
Of course, it’s humans who are actually making the decisions, but the ALADDIN algorithms would help out by making sure that the co-ordinators have the right information at the right time. ‘It’s very hard for humans to keep a handle on what’s going on in situations such as this,’ said Jennings. ‘We’re looking at getting a lot of information together from the environment, with sensors looking at what’s going on, to provide support for co-ordinate and asset allocation that is going to be done by people. The system would tell them what needs to go where, what needs to arrive together and how many units of each type are needed, and would work across agencies. Some of these agents could even be people, taking pictures on their mobile phones and sending them to news services and suchlike; crowdsourcing is something we’re definitely interested in.’
There are also two sub-projects – one looking at the evacuation of a large but confined space, for example a skyscraper or a large ship, in an emergency, and another taking into account situational awareness. ‘We’ve got a whole lot of sensors set up looking at weather conditions in the Solent area, and we’re using that to make short-term and longer-term predictions of the weather,’ said Jennings.
The consortium has expanded since Argus 2, with Bristol University bringing expertise in game theory and Imperial College adding its distributed systems know-how to Southampton’s. ‘The university teams have split into two broad groups: those that want to develop smarter agents that can reason around things and carry out planning and decision making; and the multi-agent grouping that is working out how to co-operate and co-ordinate to get their jobs done.’
Co-ordination is also a very important part of the project, with such a large team. ‘The way we’ve done it is to make sure that everyone is working on these common demonstrator projects: everyone developing all these different algorithms and techniques have to eventually put them into these common platforms,’ he said. ‘That means that you can’t go away and just develop your algorithm for just one per cent of the problem. You must take a look at the bigger picture.’
Some of the results are already in use. BAE Systems, for example, is looking at the evacuation scenario in the design of the new aircraft carriers to identify potential bottlenecks and thus ensure that the ships are as easy to evacuate as possible.
There has also been interest in the situational-awareness scenario from the UK Americas Cup team. In competition sailing, how the boats start has a huge influence on the race, and skippers are allowed access to weather information from support vessels up to a certain point, then they have to make a decision on what position to take for the start. ‘ALADDIN systems can help with that by using the available information to make a prediction of the weather conditions in the 10 minutes leading up to the start, and that could give the team a vital edge,’ said Jennings.
Aerospace & defence
The other shortlisted candidates in this category were:
ADVANCED COMPUTER MODELS FOR COMPRESSIBLE TURBULENT MIXING
This project aims to simulate the mixing of fluids caused by shockwaves, which is the process that initiates a nuclear fusion explosion. With applications in nuclear weapons, research into fusion for energy generation and in medical techniques using shockwaves to treat cancer and kidney stones, the research will further understanding of turbulence under extreme conditions.
COLLABORATION FOR SPACE EXPLORATION
Magna Parva/Liverpool University
Europe’s ExoMars mission, scheduled for 2018, will attempt to find evidence that life exists, or has existed, on Mars. Among the payload of its rover vehicle will be the fruit of this project — a lab in a box, including automated soil sample preparation, which will detect organic molecules that indicate the presence of life; if these compounds are found, they will cause fluorescent chemical reactions that can be detected by optical sensors.