Researchers study use of AED-carrying drones to save lives
By DRONELIFE Feature Editor Jim Magill
In responding to medical emergencies, such as when a person is suffering a heart attack, seconds count. Shortening the length of time between when a call is placed to a 911 operator and when medical help is delivered to the scene could mean the difference between life and death.
To meet this challenge, researchers in the U.S. and Canada are studying the development of systems that use drones to deliver automatic electronic defibrillators (AEDs), portable and easy to operate devices that ordinary people on the scene can use to keep the patient alive while waiting for the emergency medical technicians to arrive by ambulance.
A recent study by scientists at the University of Southern California used artificial intelligence (AI) and machine learning (ML) technology to examine the optimal strategies for siting locations for drone bases or depots, to ensure that the AED-equipped UAVs provide the best time savings for their emergency response.
“Drone depots, like many other things, have a little bit of NIMBYism attached to them,” Vishal Gupta, lead author of the USC study said in an interview. The USC scientists based their research in large part on an earlier study of using drones as first responders, performed by scientists at the University of Toronto.
Unlike the UT research, which focused largely on the number of bases and drones needed to achieve the best time savings across a wide geographic area, the USC scientists examined strategies for determining the best places to site the drone depots, based on limited or “noisy” data sets.
“The question that we wanted to look at was not how many drones, because we could always buy a few more, but rather, where should we locate the drone depots?” Gupta, associate professor of data sciences and operations at the USC Marshall School of Business said.
“Everyone likes the idea of this pilot program, and of using drones to save lives. Nobody actually wants a drone depot in their backyard. So, the positioning of these drone depots and the number of drone depots needed to make this system work seemed like a more first-order question in our mind,” said Gupta.
The decision on where to site a drone depot in an urban area is a relatively easy one: place the depot in a centralized location where the drone carrying the lifesaving equipment is most apt to serve the greatest number of people in the shortest time frame. The tricky part lies in how to best site drone depots in rural areas, where ambulances might have to travel long distances over gravel or dirt roads to get to remote areas and reach the patient.
Decision makers deciding where to locate drone depots in such areas often have to operate on incomplete data as to the average time it takes for an ambulance to travel to such remote locations. The data they do have are further subjected to “noise” such as how the condition of rural roads might affect an ambulance’s travel time.
“Are these dirt roads still maintained? Will they be able to find this location in this rural place?” Gupta said “Because we’re tailored to deal with that, we do a much better job of predicting travel time for rural locations, and so consequently propose drone depot locations that better serve rural communities.”
To build their model, the USC researchers used the UT research data about the frequency and location of where cardiac arrest events were happening in the surrounding Toronto area. Using that historic data and other relevant data for a given area — such as population density, the population’s median age and income level — Gupta’s team built machine learning (ML) models to predict the frequency and location of where cardiac arrest events were most likely to happen.
“I think the big contribution is the optimization algorithm. We developed an AI method that optimizes the placement of the depots and given that information, we try to make sure that we can serve the most people effectively with these drones,” he said.
The UT study had determined that for specific large region of eastern Canada, a drone-as-first-responder system would require 81 bases and 100 AED-delivery drones to reduce the average 911 response time for a cardiac emergency by three minutes.
“Cardiac arrest is one of the leading causes of death. Heart disease kills somewhere between 300,000 and 400,000 people in North America every year,” Justin Boutilier, the lead author of the UT study, said in an interview. “In general, we find that drones can, of course, improve response times and you don’t need a large number of them to do it.”
Boutilier, who co-authored the study as a PhD student at the University of Toronto and is now an assistant professor at the University of Wisconsin Madison, said there are a number of pilot programs for using drones to deliver AEDs to cardiac patients underway in the Toronto area and in Salt Spring Island, British Columbia on the west coast area of Canada.
“There have been some tests in the U.S. as well. There’s a group at Duke that’s doing research on this topic, and has been collaborating with the EMS [emergency medical service] folks.” Several cities in Sweden have already implemented such drone response programs and in January officials there for the first time credited a drone-delivered AED with saving someone’s life.
“I think something like this needs to happen in the cardiac arrest space, especially with out-of-hospital cardiac arrest, so that we can actually see improvement in outcomes here,” Boutilier said. “I’m excited about it.
Miriam McNabb is the Editor-in-Chief of DRONELIFE and CEO of JobForDrones, a professional drone services marketplace, and a fascinated observer of the emerging drone industry and the regulatory environment for drones. Miriam has penned over 3,000 articles focused on the commercial drone space and is an international speaker and recognized figure in the industry. Miriam has a degree from the University of Chicago and over 20 years of experience in high tech sales and marketing for new technologies.
For drone industry consulting or writing, Email Miriam.