Research News

UB engineer sees drones as more than eyes in the sky for disaster response

(L-R) Souma Chowdhury, assistant professor of mechanical and aerospace engineering in the University at Buffalo’s School of Engineering and Applied Sciences, together with ADAMS Lab research associates Sanchit Gupta, Arun Sunil and Maulik Kumar demonstrated how collaboration between autonomous drones can provide communications and other assistance to emergency responders.

(From left) Engineering professor Souma Chowdhury and ADAMS Lab research associates Sanchit Gupta, Arun Sunil and Maulik Kumar demonstrate how collaboration between autonomous drones can provide communications and other assistance to emergency responders. The demonstration took place during the full-scale emergency exercise conducted July 19 on the South Campus. Photo: Meredith Forrest Kulwicki

By MICHAEL ANDREI

Published July 27, 2017 This content is archived.

Print
headshot of Souma Chowdhury.
“UAVs are increasingly coming to the forefront of humanitarian technology innovation. ”
Souma Chowdhury, assistant professor
Department of Mechanical and Aerospace Engineering

Within hours after a 7.8 magnitude earthquake struck Nepal in April 2015 killing and injuring thousands of people and destroying thousands of homes, foreign governments and aid agencies began sending relief workers, medical equipment, food and clean water supplies to the Nepalese government.

The following year, international aid agencies sent food rations, and medical and relief supplies to Ecuador after an equally strong earthquake struck in April 2016.

And government workers and emergency responders in both disasters also received something more: drones.

“In Nepal, unmanned aerial vehicles — UAVs — were used for search and rescue missions and to map out buildings, homes and other structures destroyed by the quake,” says Souma Chowdhury, assistant professor of mechanical and aerospace engineering, School of Engineering and Applied Sciences.

“UAVs were also useful in evaluating infrastructure, including roads, that were damaged in the Ecuador earthquake to help determine whether they were safe to use or not.”

Chowdhury sees UAV technology playing an increasingly critical role in emergency responses and disaster-relief operations around the world.

“UAVs are increasingly coming to the forefront of humanitarian technology innovation,” he explains. “While their application in disaster-relief situations is still in the embryonic stage, recent tests have yielded encouraging results and are also demonstrating the versatility of the drone as a platform for other technologies.”

Some of those results are coming from Chowdhury’s Adaptive Design Algorithms, Models and Systems (ADAMS Lab) laboratory.

His research pioneered a way to program teams of UAVs — swarming drones — to quickly and efficiently map out oil spills.

(L-R) Souma Chowdhury, assistant professor of mechanical and aerospace engineering in the University at Buffalo’s School of Engineering and Applied Sciences, and ADAMS Lab research associates Sanchit Gupta and Arun Sunil demonstrated collaborative UAV flight during the full-scale emergency exercise conducted July 19 on the UB South Campus by the University at Buffalo Emergency Management department.

(From left) Engineering professor Souma Chowdhury and ADAMS Lab research associates Sanchit Gupta and Arun Sunil demonstrate collaborative UAV flight during the full-scale emergency exercise conducted July 19 on the South Campus. Photo: Meredith Forrest Kulwicki

Drone in flight.

A drone in flight. Photo: Meredith Forrest Kulwicki

Chowdhury devised a method for UAVs to quickly record whether they are over water, oil or the edge of the spill. With multiple drones making observations every five seconds, the size of the spill can be determined quickly. Moving point to point over the spill, they avoid going over space previously covered by other drones.

UAVs’ applications for emergency and disaster response globally and in the U.S., Chowdhury says, can vary from mapping an environmental hazard to saving lives during such events as hurricanes, wildfires, tornadoes and floods by detecting victims or transporting critical payload — including medical or communication supplies — into hard-to-reach regions.  

“Which means you need information over a large area … ideally from the air,” he says.

“In the last five years or so, UAVs have become pretty state-of-the art. We can buy very stable platforms for a couple thousand dollars right off the shelf.

“The only difference is that if you employ only one drone, it is not going to be enough. You need many more than that because, sure, you can fly one UAV over an area to collect data,” Chowdhury says.

“But time is of the essence when you are trying to save lives. These are what we call dynamic events, or time-sensitive events … you need to maximize the amount of information gained in a minimum amount of time.

“So, now I need to put multiple eyes in the sky … which is what swarm technology is about,” he says. “If you want to get those multiple eyes up and have them optimally collaborate with each other … the question becomes, ‘how do you do that?’”

To do this, Chowdhury says, the drones must be able to manage themselves.

“If a single person is required to manage each UAV, then there will be emergency responders, members of your fire department or police department, who will be needed simply to monitor or fly the UAVs instead of doing something else, something where skilled human operators are needed more,” he says.

“Ideally, you would want these UAVs to fly themselves, avoid collisions with buildings and with each other, figure out where to go next, how best to execute the mission as a team, and so on.

“So, now we are talking autonomous, and collectively autonomous, drones. This is where things start to get more difficult, especially when you want to accomplish such a level of sophistication with a frugal budget, typical of civilian small UAV applications” he says.

Chowdhury, however, envisions teams of heterogeneous UAVs in emergency and disaster response.

“To actualize this, some UAVs might be collecting imagery data, others taking infrared readings — air quality readings in a fire or hazmat event — and a third set of UAVs in that team could be carrying a payload of medical supplies or communications equipment to be dropped off at critical locations as identified by their imaging teammates,” he says.

“There is also growing interest in designating a UAV that would serve as a communication drone, and through that UAV, the others could talk to each other — thus creating what we call a flying ad hoc network, or FANET,” he says.

“Each drone might be performing a different task, but also collaborating with each other, all on the same team.”

Chowdhury says his research is focused on networking UAVs, providing them with the technology to talk to each other, make decisions and move where they are needed without someone on the ground having to synchronize the activity. Maintaining wireless ground communication with the swarm could be done by a human being, he says, or a robot that is communicating with the control node UAV.

“Another key point here is that you want to be able to deliver medical supplies or communications equipment as rapidly and efficiently as possible while the disaster is ongoing,” he says.

“This is a big advantage of UAVs — they can go quickly and easily in areas that might not be accessible or practicable for a helicopter, for example.”

Chowdhury notes there is ongoing research focused on resolving open questions regarding UAV technology. “We are focused on taking this technology where it can provide even greater assistance,” he adds.

“In a flood, for example, if there are people stranded on top of a house, the UAV can take images, recognize them as human beings and communicate with other drones, then tell them to go and drop a package there, such as food rations or first-aid supplies.

“As the versatility of UAVs increases, and extreme weather becomes more severe, they will play a larger role in disaster monitoring and response at different scales.”