The role of Robot Dogs and Drones in Search and Rescue operations

[berkeman]

This was a great use of technology last week when a large parking garage collapsed in New York City, and it was too dangerous to send in FireFighters to search for injured and trapped people. FDNY has used Drones for a while now, and has a new Robot Dog that is able to navigate pretty difficult rubble to search for survivors:

https://www.firehouse.com/tech-comm...ised-for-searching-collapsed-garage-last-week

Last week, when a parking garage collapsed on Ann St. in lower Manhattan, the FDNY deployed its so-called Digidog to search for people trapped in the wreckage and survey the scene — a move Adams touted at the time because it eliminated the need to put firefighters in harm’s way. The robot and its handlers were engaged in a training exercise at a downtown high-rise when the garage came down and arrived about 15 minutes after the first distress call went out, Kavanagh said.

“The robotics members went to work quickly, suggesting that the robotic dog could aid in the operation,” she said. “Three drones were also deployed — one overhead to give a clear picture of what was happening on the roof and two inside to assess the situation in there.”

Adams noted that it was the first time in the city’s history that the FDNY and NYPD responded with drones and the Digidog, which is named Bergh. Fabien Levy, Adams’ spokesman, said the FDNY currently has one Digidog and the NYPD has ordered two.

“This is an administration that is not going to be fearful of using everything possible to save the lives of New Yorkers and to save the lives of first responders,” Adams said. “The benefits are not theoretical. Last week, we saw it in action.”

 

[jedishrfu]

My question is how good are they at finding people? Will they miss more people than a human or real dog searcher due to tech limitations?

It kind of brings back the opening scene from Wil Smiths I Robot where the Robot rescuer saves him over the young girl based on a rational analysis of the situation.

 

[berkeman]

My question is how good are they at finding people? Will they miss more people than a human or real dog searcher due to tech limitations?

AFAIK, they only have visual and auditory search capability, so no scent search capability. Yet.

EDIT -- Although maybe they have Bluetooth search capability, and can search for cell phones to locate people. That would be a logical feature to include. I'll have to mention that to my cap'.

 

[FactChecker]

AFAIK, they only have visual and auditory search capability, so no scent search capability. Yet.

EIDT -- Although maybe they have Bluetooth search capability, and can search for cell phones to locate people. That would be a logical feature to include. I'll have to mention that to my cap'.

After AFAIK, I searched the internet for the meaning of EIDT without success. Then I realized that it probably was literally EDIT.

 

[berkeman]

Doh! Fixing that now, thanks.

 

[berkeman]

My question is how good are they at finding people? Will they miss more people than a human or real dog searcher due to tech limitations?

Paging our own SAR dog trainer specialist @DaveE

 

[FactChecker]

Although maybe they have Bluetooth search capability, and can search for cell phones to locate people. That would be a logical feature to include. I'll have to mention that to my cap'.

Good thought! Here is a related effort: Bluetooth RSSI locator robots

 

[256bits]

My question is how good are they at finding people? Will they miss more people than a human or real dog searcher due to tech limitations?

Digidog doesn't really do much of anything, except that it can be used in situations like this where building integrity is questionable. Attach a heat sensor, thermal camera, listening device, or any other sensor to Digidog, move it into the unsafe area. The human operator would receive information to be conveyed to the actual firefighters to aid in rescue by pinpointing areas of search. The operator I would think would have to guide the dog through the rubble.
Perhaps future versions could be adapted to be self sufficient in that regard, with a self contained AI designating search patterns and monitoring its own progress.
IT is really not a very smart dog.

 

[FactChecker]

The human operator would receive information to be conveyed to the actual firefighters to aid in rescue by pinpointing areas of search. The operator I would think would have to guide the dog through the rubble.

This seems like the smart way to make immediate progress. I don't think it will be long before we can use a swarm of quadcopter drones that use AI to conduct a search. The future is promising.

 

[Klystron]

I read in newspaper articles, not peer reviewed publications, that walking robots and airborne drones used CO₂ detectors to locate humans and pets buried in rubble by tracking their exhalations.

One could speculate that robotic platforms such as purchased by the NYFD and NYPD be programmed to track a signal from any portable electronic detector such as for chemical explosives, smoke, gamma radiation, land mines, RF emitters such as communication gear or active embedded medical devices, or weak auditory signals such as crying. Multiple platforms with multiple detectors also provides signal triangulation and position.

Modular theory suggests that a general purpose robotic dog equipped with mission specific detector modules would be more versatile and cost effective than a robot designed for a specific task, say bomb detection.

 

[jack action]

My question is how good are they at finding people? Will they miss more people than a human or real dog searcher due to tech limitations?

It kind of brings back the opening scene from Wil Smiths I Robot where the Robot rescuer saves him over the young girl based on a rational analysis of the situation.

This is the problem when people mix reality with fiction.

These machines do not save people, they assist people who do. Obviously, these machines do not decide who gets to be saved; they just identify who needs to be saved, give their locations and report on the conditions of the surroundings. Based on that, humans still make the decisions on what to do and they still go in do the saving themselves.

 

[jedishrfu]

I wasn’t mixing reality and fantasy. I was thinking of the myopic view an operator has in using remote control robotics where you can only see what the camera sees and may miss an important cue.

 

[berkeman]

These machines do not save people, they assist people who do. Obviously, these machines do not decide who gets to be saved; they just identify who needs to be saved, give their locations and report on the conditions of the surroundings. Based on that, humans still make the decisions on what to do and they still go in do the saving themselves.

Agreed. In this incident, the collapsed structure was unstable, so it was not safe for FFs to conduct normal SAR operations. If the Digidog or a drone had identified a trapped/injured person, the operation would have been able to focus on that location for a shoring rescue or worst case a helo hoist rescue. More tools in the operation is a good thing, IMO.

 

[DaveE]

IT is really not a very smart dog.

Yup. Wait... it's just not a dog at all. This is a robot that carries sensors and sends data to humans. Note that none of those sensors are olfaction; which, BTW, is the ONLY reason to have a real search dog.

Also, notice the format of the news stories which, with a bit of editorial bias on my part, goes like this: we bought a glitzy high tech thing; a building collapsed and we sent that thing in to help; end of story. So what was achieved? Did the robot save anyone? How would a fire department that spent that money on something else have faired? Maybe it's good technology, or will be someday, but you'd never know from this PR effort.

The whole "dog" part of this is a sympathy ploy because people like dogs, dogs are known for helping in similar situations (although they do a different job), and everyone involved wants to generate interest. Sorry for being the curmudgeon, but there's no real story here AFAIK.

I think remotely deployable sensors are a great technology to pursue, but let's leave the whole "dog" thing out of it. Sometimes you want a dog, sometimes you want a robot. How about covering how the robot did good robot things?

 

[FactChecker]

Yup. Wait... it's just not a dog at all. This is a robot that carries sensors and sends data to humans. Note that none of those sensors are olfaction; which, BTW, is the ONLY reason to have a real search dog.

I really associate the "dog" reference with 4-legged locomotion as opposed to wheeled, tracked, or flying. I think there are real advantages to the 4-legged method, especially when there are heavy weights to be carried over rugged terrain. Flying drones have a great advantage if the weights are light enough and long endurance is not needed.
I am sure that people with more expertise than I can elaborate more.

 

[Joshua Akano]

I really associate the "dog" reference with 4-legged locomotion as opposed to wheeled, tracked, or flying. I think there are real advantages to the 4-legged method, especially when there are heavy weights to be carried over rugged terrain.

You have a good point, FactChecker, but even with that, I'm thinking that the "dog" might not possibly make it through a sudden collapse caused by the continuous burning of the building.
A good point also to consider is its speed (flying could help increase this, although weights slow it down) and toughness (toughness doesn't have relation with the legs but still needs to be pointed out).