Technology based on Artificial Intelligence is helping keep people safe during an emergency.
New software called Incident identifies firefighters from victims and sends information to first respondents to help speed up the rescue process.
EVisuals developed the software with the University of Wollongong’s Smart Infrastructure Facility. EVisual’s managing director Matt Lynch has repeatedly seen tragic outcomes due to trained fire wardens feeling overwhelmed during an emergency evacuation.
“Using Nvidia hardware and artificial intelligence, Incident can tell the difference between firefighters searching the building from other building occupants. It can also identify any people who need help to evacuate, due to a disability, for example,” he said.
“This is an ambitious project, but with the help of SMART, we’re confident that the system can make an incredible improvement in building safety and emergency incident control.
“Confirming the location of building occupants is currently reliant on verbal reports from building wardens and extensive searches by emergency services.
“Incident gives emergency responders the critical information that they require about the building, and its occupants, in real time.”
SMART researcher Dr Johan Barthelemy says the system will be able to locate building occupants, differentiate them from firefighters, and has the potential to save lives in a range of scenarios.
“The idea for this project is to train an artificial intelligence model that will be able to recognise people, wardens and firefighters,” he said.
“The system can be used wherever fire wardens are currently appointed. That means in office blocks, residential apartment blocks and industrial complexes, for example.”
SMART director, Senior Professor Pascal Perez, said the facility was chosen for the six-month project because of the level of expertise on offer.
“This project builds upon and improves our current Edge computing technology, in order to adapt it to in-building emergency evacuation,” he said.
Next stage for the project is identify a test site.
“We hope this technology-enabled project will help emergency services organising better and faster evacuation during various disaster situations,” said Senior Professor Perez.