When firefighters need to enter smoke-filled buildings to conduct search or rescue, they frequently suffer from low visibility and often need to feel their way along walls or follow ropes reeled out by the lead firefighter. With a limited supply of oxygen carried by each firefighter, being slowed by the inability to see can severely limit their capacity to carry out duties in these environments. Now researchers from King’s College London and Sheffield Hallam University have developed a prototype robot assistant for firefighters that can help guide them through even the thickest smoke.
Rather than an independent robot firefighter in itself, this device is instead intended to act somewhat like a guide dog where it leads a firefighter into a rescue area, with the firefighter following on behind holding onto the robot's "reins." According to the researchers, this allows the firefighter to move rapidly through hazardous, low-visibility conditions, as vibrations transmitted through the reins alert the user about the size and shape of any obstacle in its path.
Apparently it can even provide other information about encountered objects, such as their pliability or stiffness, which may also help the user locate unseen unconscious people.
Equipped with tactile sensors, the new guiding robot is designed to provide haptic feedback to the user being guided that indicates various types of objects. In a real-world situation, according to the researchers, the firefighter would don a special sleeve that incorporates electronic micro-vibrators along its length. Enclosing the whole arm, the sleeve is designed to provide feedback that the firefighter would have been trained to interpret from the signals sent back by the robot.
The team claims that the robot guider prototype is designed to respond to any change in the firefighter's reactions too, such as if there is hesitation or resistance to forward movement. When this occurs, the guiding robot slows down to match the pace of its user.
The robot is also designed to preempt or predict the user's actions based on an interpretation of their movements as well as a catalog of their previous actions. According to the researchers, this process is controllable via an algorithm they have created that could determine the firefighter's degree of trust in the machine, so that it can tailor its interactions accordingly.
"We’ve made important advances in understanding robot-human interactions and applied these to a classic life-or-death emergency scenario where literally every second counts. Robots on reins could add an invaluable extra dimension to firefighting capabilities," said Dr. Thrishantha Nanayakkara of King’s College London.
The tactile language being developed by the team is also being considered for use by robotics in a range of domestic settings, and the researchers now intend to further investigate how reins and haptic controls may help the elderly be more independent in their own homes.
Before such advanced uses come to pass, however, the team must first prove its device in more real-world scenarios.
So far the guiding robot has been tested in smoke-filled caves in Germany, going past objects and potential hazards while guiding blindfolded test subjects around. Now that a proof-of-concept has been completed, the team plans to build a full working prototype for testing in real-world firefighting conditions.
"With the use of robots in emergency situations still in its relative infancy, it is crucial to develop an understanding of how robotics interact with people and how those communications can be explored," said senior designer Heath Reed of Sheffield Hallam University.
With project partners that have included the Guide Dogs charity, South Yorkshire Fire & Rescue Service, and Thales Ltd, the team is hopeful of continued research and eventual applications that improve the haptic feedback system and its use in interactions between robot assistance and humans.
"This project paves the way for robotics to be developed in a number of exciting sectors and I would expect the next five years to see some real developments based on our own research," said Jacques Penders from Sheffield Hallam University.
This research has been developed with funding from the Engineering and Physical Sciences Research Council (EPSRC).
Source: EPSRC
What problem is this device trying to solve? The article mentions these mysterious obstacles without really going into detail. Bumping into a wall or a chair won't kill you - falling through the floor will; a Cincinnati firefighter was just killed by falling into an open elevator shaft. I don't know what led the researchers into thinking that firefighters are dying because they're going too slow while searching. In pure statistics, most firefighters do not die while actually in a structure fire. However, if you want to focus on navigating a fire building, look at the worst problems: falling through the floor and getting lost and running out of air. See what can be done to supplement good search practices. Keeping an axe handle or Halligan out in front of you, sweeping the floor and sounding stairs will detect almost all obstacles. What can we come up with that will find anything that these practices will miss?
"This device is instead intended to act somewhat like a guide dog where it leads a firefighter into a rescue area..." Again, what exactly is it doing? Is it taking the firefighters into the building or out of it? If the robot isn't trying to find a way around an object, then it's not doing anything except getting in the way. In most structures, two firefighters can't fit side by side, if the robot decides it needs to make a U-turn there probably won't be room for it to get by. And any robot that is going to fulfill these duties will need to be fairly large in order to get over any of these obstacles and at the same time be robust enough to stand up to fire conditions. How about a little car that I can roll down the hallway that will beep if it detects that it's falling? Bonus points for switching the design to anything that isn't tethered to the firefighter.
"According to the researchers, this allows the firefighter to move rapidly through hazardous, low-visibility conditions..." On first inspection, this seems like a good goal, but let's think about this. You would never tell a Probie, "Go as fast as you can." Better advice would be, "Go slow so you have time to react if there's suddenly no floor in front of you. Go slow so that you can tell the difference between a victim and a couch (or a victim on a couch)." Search and Rescue is not about racing through the smoke, it's about finding something of value; the location of the fire, or any victims (or the lack thereof). Improvements to Search and Rescue should not focus on speed, but rather thoroughness. You don't want another Company to come and say, "We found a body that you missed on the primary search." When the adrenaline is pumping, going too slow is not usually a problem.
"In a real-world situation, according to the researchers, the firefighter would don a special sleeve that incorporates electronic micro-vibrators along its length." Haptic feedback is a unique way of communicating information, but it requires the ability to concentrate on the feedback. I would encourage the researchers to put on turnout gear and an air pack, grab a set of irons and go into a smokehouse. Then, start to perform a primary search while trying to identify which of the 7 straps on your arm is vibrating. A helmet mounted display would be much more easy to use; the facepiece-mounted LED indicators for air packs are simple and reliable. Why not try and optimize this area into a small dashboard of information?
Lastly, when this is in your first paragraph, "supply of oxygen carried by each firefighter," I start to question how well you know the subject matter.
I believe that the reason for wall following is that when room is smoke filled and the floor below is aflame is that the are the wall provides the best chance of avoiding a fall through the floor due to weakened joists.