A dog’s snake robot called CAT (seriously)

In Features /

Ryerson’s Network-Centric Applied Research Team had the chance to test out out their emergency response technology last weekend. Story and photos by Online Editor Jeff Lagerquist

It’s 8 a.m. A blue sedan travelling south on Hurontario Street in Mississauga is pulled over by Peel Regional Police for a routine traffic stop. The driver is visibly nervous. As he frantically searches the glove compartment for proof of insurance, the officer detects a pungent chemical odour coming from inside the car. He detains the driver at the side of the road as his partner rifles through a backpack laying on the back seat. Inside, they find large containers of chlorine and brake fluid, common household items used to build bombs.

The officers complete their arrest as the driver sits cuffed in the back of the cruiser. Suddenly the radio crackles and the dispatcher demands that all available officers respond. A bomb has exploded at the City Centre Transit Terminal at the Square One Shopping Centre. Hundreds of commuters catch Mississauga and GO buses here every day. The explosion tore apart one bus and damaged many others. Dozens of people are trapped under the collapsed terminal building.

Police dispatch the urban search and rescue (USAR) and the Chemical, Biological, Radiological, Nuclear and Explosive (CBRNE) response teams. The building could collapse at any moment, and they suspect there may be deadly chemicals contaminating the area. Ryerson’s Network-Centric Applied Research Team (NCART) is briefed and called to the scene.

Alex Ferworn is the graduate program chair and associate chair of the department of computer science. His lab developed a disaster scene reconstruction program that provides rescue professionals with a 3D model when it may not be safe for humans to enter. His team also works with rescue dogs, fitting them with cameras and deployable supplies to help victims survive under a pile of rubble.

Three of Ferworn’s master’s students and a PhD researcher casually leave the George Vari Engineering and Computing Centre around noon. They load the car, and fumble with an iPhone for directions once the journey is underway. The students agree that a trip to McDonalds is in the cards. The leisurely departure from campus is no big deal. In reality, there was no bomb, no explosion, and no suspicious vehicle emanating a chemical odour on March 4 — just a milliondollar simulation scenario paid for by Public Safety Canada.

The multi-agency operation, hosted by the Ontario Provincial Police (OPP) at the USAR and CBRNE (UCRT) Centre for Excellence outside Bolton, Ont., was kept under wraps until a few days beforehand. Ferworn and his students were invited four days ago to test and showcase their equipment. The Bolton scenario replicates the aftermath of a bomb attack at a busy bus terminal.

An old TTC bus blocks the entrance to a concrete room buried beneath a huge pile of concrete rubble, and another bus sticks out of the pile at a 45-degree angle. Inside the buildings and buses are mock victims with simulated injuries. The ceiling and wall can be adjusted with cables to simulate various degrees of collapse. The UCRT team begins bracing the outer wall of the building, a process they call “shoring,” so the rescue effort can begin.

Jimmy Tran is three years into a PhD in computer science at Ryerson. He starts to warm up the hexrotor Unmanned Aerial Vehicle (UAV) that will carry the camera equipment over the disaster zone.

“The camera is sensitive to light. It’s meant to be used indoors,” says Tran. The 3D camera used by the NCART team is actually an Xbox 360 Kinect camera. “We just stripped the casing and motor out of it to save weight.”

Sunset provides the ideal conditions for the Kinect camera. There is just enough light for the colour camera to see but not enough to prevent the infrared camera from measuring depth. Advanced cameras able to record in bright sun or darkness would cost over $10,000 and require a considerably larger UAV. The Kinect, however, costs under $150.

The data gathered by the camera generates a 3D point cloud — a series of measurements that correspond to the topography of the terrain.

“My students use the 3D point cloud to generate a photo mosaic of the area. You can rotate it 360 degrees and look for people and potential hazards. It allows rescuers to see what’s going on and map out a plan without actually having to be in there,” says Ferworn. “We call this ‘stand off.’ ‘Stand off’ is very important when you don’t want to be killed.”

Tran and the team huddle around the UAV and a laptop, set on a folding table near the rubble pile. However, the UAV struggles to get off the ground several times as the batteries struggle to endure through the cold weather. They attach a lighter camera and a short flight over the rubble pile successfully gathers some 2D video. Ferworn calls it a wrap, and the team retires to their mobile command centre on loan from the OPP.

The interior of the massive vehicle provides enough workspace to tinker with the UAV and enjoy a military ration meal. The UCRT team is equipped to work, eat and sleep at the scene of a disaster for several days. An OPP officer boards the command center during dinner. The chatter on his walkie-talkie calls for the canine augmentation technology (CAT). USAR dogs wear CAT as they search for the survivors of an urban disaster.

A vest designed by the Ryerson School of Fashion holds cameras mounted on either side of the rescue dog allowing their handlers to see from the dog’s unique perspective. The system has evolved since its inception in 2005. CAT now features the canine remote deployment system (CRDS), which carries emergency supplies including a radio, food, and water in a bag below the chest called an Underdog.

“The CRDS listens for the dog’s bark and releases the Underdog bag when he signals a found victim,” says Ferworn. But the Ryerson-patented technology was not without its problems. “We discovered that radios don’t work very well in rubble,” he says.

To help with this problem, NCART teamed up with Carnegie Mellon University in Michigan to pair the CAT and CRDS systems with a robotic snake last November. The snake coils into an “S” to fit under the dog. Like the Underdog survival pack, it deploys with the dog’s bark and can reach deep into a cavernous rubble pile.

“Nobody had ever tried to deploy a robotic snake from a dog before,” says Ferworn. “We did, and it worked. It went in 60 feet and sent images back. It was a huge success.”

The snake is built from eight aluminum segments and a socklike fabric acts as a kind of skin. Its only weakness is the half-inch wide cable that connects the snake to the operator. At the Bolton site, OPP officer Dan Bailey deploys Dare, an 11-year-old black Labrador wearing the CAT system.

He finds a victim in seconds, and the CAT system confirms a safe route for the UCRT team to make the rescue. Dare is one of only two Federal Emergency Management Agency (FEMA) dogs in Canada. Determined to prove the merits of their research, the Ryerson team abandons the UAV for the evening. Second-year computer science master’s student Alex Ufkes walks with a laptop behind Tran as he holds the Kinect camera in his hands.

They make a pass around the bus and the collapsed building entrance, recording a 3D point cloud under the powerful spotlights. Ufkes eats peanut butter from a foil ration pack as he sits at his laptop inside the mobile command centre. The software processes the images and develops a 3D rendering of their pass. Ferworn stands over his shoulder. Even though a flight over the rubble pile would have produced a more impressive model, he seems pleased.

“No team in Canada has this kind of capability,” he says.

CAT and 3D reconstructions of disaster environments have the potential to save lives and aid rescue workers in extreme emergency situations like the Fukushima nuclear meltdown in Japan and the 2011 earthquake in Haiti. A high-ranking OPP officer enters the mobile command center around 1:45 a.m. He stares at the rotating 3D model of the broken concrete and twisted metal outside.

All he can say is, “That’s sweet, eh?”

Leave a Comment