The relation between technology and safety is twofold: on one side, new technologies can be the cause of emergencies (for example, a bug in the control software of an airliner); on the other side, technology can provide tools to better plan, prepare for, and handle emergencies. I've spoken about the latter with Chris Johnson, international expert on safety and professor at the University of Glasgow, Scotland. Chris works with many organizations (for example, NASA, Eurocontrol, health and security authorities in the UK,…) and is in Italy this week. Here's what he told me:
What does recent research tell us about human behaviour in emergencies?
During emergencies, we have observed a range of different human behaviours that are often determined by the context in which the emergency occurs. This can best be illustrated by two aircraft fires. In the first example, there were many fatalities as families rushed to escape the flames and smoke – trying to save their loved ones. In contrast, a comparable accident on a commuter flight led to a small number of minor injuries because the passengers did not panic or rush to escape. In these two cases, the desire to save a loved one led to significant differences in both the behaviour of the passengers and the outcomes of the accident. Similarly, if we look at fires in night clubs, the presence of alcohol combined with limited knowledge about the location of fire escapes often leads to higher levels of casualties than would otherwise be expected for instance in similar fires in everyday working environments such as shops or offices.
In spite of these differences, there are also some common features between different emergencies such as fires or the structural collapse of buildings. For example, there is a deep seated tendency for people to go out the way that they came into a building even if this means that they walk past emergency exits. These doors often have signs saying ‘not to be used unless in an emergency’. People also often have concerns that emergency exits will be locked. They may not be sure where the emergency exits will take them. This lack of familiarity with emergency egress routes continues to cost lives.
How can technology help us to respond to these emergencies?
One of the reasons for my visit to Italy is to see the ‘serious games’ work that the team at the HCI Lab in Udine have been developing. These systems enable people to simulate the evacuation of buildings under different emergency scenarios that would not be possible without holding many different drills or exercises. There has also been work on using mobile devices that provide building occupants with up to date information on the best routes out of a building. In Edinburgh, another team has integrated these devices with sensors that are embedded into the structure of a building. This is important because in the future these sensors will tell the emergency services and the building occupants about whether or not it is safe to go into particular areas of a building following a fire or an earth quake. This advice can be updated from minute to minute as the sensors detect changes in the structural integrity of the building.
My own work looks more at the design of evacuation procedures for environments where it is difficult or impossible to hold drills and exercises. These include the cardiology wards of hospitals (see figure below) where the patients are too ill and the hospital units are too important for them to be closed in order to practice an evacuation. We have also looked at football stadiums where we use computers to model the egress of up to 50,000 people into the streets surrounding the sports venue. In this case, the costs and also the ethical concerns about injuries in the crowd mean that we cannot easily hold practice evacuations of the entire stadium. We have also modeled the evacuation of the underground train network in Glasgow. In this situation, it is possible to hold exercises with real people. However, these have to be held very late at night when the train system is not being used to carry fare paying passengers. The operators also cannot take the risk of using children or the elderly in their planning exercises. In contrast, we can program our computers to model what might happen when family groups take the same actions to escape with their loved ones as was described for the aircraft example in the opening paragraph.
Do you have any specific examples to illustrate the opportunities in the use of technology to support emergency planning?
I would highlight two particular areas that might be of interest at the moment. Earlier this month we provided copies of our software to the Integrated Security Unit for the Vancouver 2010 Winter Olympics. We are also working with groups that support the planning of the 2014 Commonwealth Games in Glasgow. Simulators such as the football stadium model mentioned above, can be used to predict what might happen when police and stewards must direct large crowds within a sporting venue. A particular benefit of software systems is that these crowd models can be analysed well before the venues are completed. This is very important for most Olympic planning committees where the stadiums and other arenas are often completed with only months or weeks to go before the competitions begin. The tight construction schedules often leave too little time for the police and security organisations to conduct a wide range of crowd control and other planning exercises. These can be done using the software simulations while the venues are still being built.
A second example of the innovative use of technology to support emergency planning is in counter terrorism. In particular, we have been building computer simulations to consider what might happen if terrorist groups were to transfer techniques that have been used in Iraq, Afghanistan and Mumbai to the streets of major European cities. For instance, there have been a number of recent attacks using coordinated detonations of Improvised Explosive Devices. A small explosion is used to trigger an evacuation – crowds form around exits, at assembly points outside a building or in the streets around a market. These crowds then become the target for secondary devices, typically carried by suicide bombers, who try to maximise the casualties among people who are fleeing from the first explosion. We can use simulators to show how the evacuation plans that help people escape from fires or from structural collapses also create the crowds that become the targets under terrorist attack. The software is interactive and so security forces can plan alternate evacuation strategies that minimise the numbers of people that are groups together following an initial attack (the screenshot below concerns a UK Rail station).
A final example is the extension of these techniques to provide epidemiological simulations of the impact of pandemics. 
0; We can expand the scope of the computer software to account for the movements of people not just within single buildings but across cities and regions. The social contacts that occur with these various movements also provide a basis for modelling the spread of infectious diseases, including the H1N1 and H5N1 variants that have been the focus of recent pandemic concerns. Government agencies can then use the predictions from these models to look at the trade-offs that might occur between different immunisation policies or the use of anti-viral agents in the case of H1N1 and H5N1, as well as physical containment strategies. The high-level aim behind all of this work is to increase civil resilience in the face of a growing number of threats.
