Computer programme simulates emergency response when a city is hit by a natural disaster

Dealing with disasters could become a lot easier thanks to a computer programme developed by a team of researchers at The University of Auckland.

The Artificial Intelligence Group in the Faculty of Science’s Computer Science department are working on the Robocup Rescue Simulation Project which aims to simulate the emergency response when a city is hit by a natural disaster.

The Robocup Rescue Simulation Project was established by Japanese researchers after the 1995 Kobe earthquake. The project simulates the effects of an earthquake and its aftermath, and the response of emergency services.

Each year a competition is held to test the abilities of various research groups around the world. Teams are given a simulated earthquake situation and scores are based on the number of civilians who survive and the area of the city that is saved.

"The competition drives the research forward," says Cameron Skinner, research programmer for the Artificial Intelligence Group.

"It gives us an opportunity to compare our work with other leading researchers from around the world, and it gives everyone an incentive to write software that works better than before."

Last year, "The Black Sheep" - the Auckland team - entered the competition for the first time and was placed fourth out of 20 teams. They were also invited to join the technical committee which guides the development of the competition and decides on the rules.

This year the team have again qualified to compete in the Robocup competition in Lisbon in late June against 20 teams from Europe, North America, Asia and the Middle East. Cameron, who is working on his PhD, and Jonathan Teutenberg, a recent Master of Science graduate, will be attending this year.

The competition involves three days of heats, followed by two days of semi-finals and finals.

“At the competition last year we looked at what others did right and what we did wrong during the day and then sat up all night rewriting our programmes to account for those factors,” says Jonathan.

There are three dimensions for the Auckland team’s research. The team is trying to develop physical simulators that match the real world as closely as possible, for example a fire simulator that accurately predicts which buildings will catch fire and when. They are also producing policies for the simulated emergency services.

“Hopefully these policies will translate into ideas that the real emergency services can use,” says Cameron.

Simulator development has been boosted this year by the introduction of an infrastructure component to the competition. Teams can develop new simulators and submit them to the competition and the best ones will be used the following year for the main event.

"Our third research direction is to take the Robocup Rescue simulator and turn it from a research project into tools that might be used by Civil Defence to help with their planning and training," says Cameron.

The group are working with Auckland Urban Search and Rescue (USAR), the fire department and the Police to develop these tools.

Computer science lecturer and a member of the Artificial Intelligence Group, Dr Mike Barley says the team hopes to be able to produce prototypes for systems that Civil Defence and USAR can use to test their existing emergency management plans, and also to help with scenario development.

"Eventually, if the simulators become realistic enough, we will be able to ask the computer what would happen if an earthquake of a certain magnitude occurred, how many lives could be saved if we had two extra fire trucks, and so on,” says Dr Barley.

The ultimate goal of the Robocup competition is to be able to create robotic vehicles for real life situations, although this is unlikely to become a reality for at least 50 years.


The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of News Medical.
Post a new comment
You might also like...
New research explores the unique effects of ADHD medications on Drosophila brain cells