Navy robot could fight most dangerous fires aboard ship
Just not for a few years, until it grow up enough to find its way around
Endless optimism in the face of obviously difficult, complex difficulties has led humans to succeed long wars, revolutions, plagues, difficult terrestrial explorations, the quest for space and the belief that despite their artificial immortality, neither Tang nor Twinkies would ultimately poison those who loved them.
The Naval Research Laboratory is pushing unrealistic optimism to new lengths with a project designed to create a humanoid firefighting robot that can survive toxic smoke and chemicals, murderous heat and cramped conditions to fight fires aboard Naval ships that would kill human firefighters.
It's a sensible, even noble intent. Fire, not explosions or gunfire was the biggest fear of sailors under attack by Kamikaze in World War II and in nearly every naval war before that, too.
Wooden ships burn easily, especially when you cover them with pitch and fill them with gunpowder. A ship that burned was a ship that sank, and most sailors couldn't swim. So ship that burned and sank would kill most of the crew who survived the disaster itself.
On metal ships the problem became more intense, as the number, length and inaccessibility of the territory belowdecks expanded in a maze of long, low, cramped hallways and unexpected caches of fuel, explosives and flammable chemicals, often under pressure.
The Shipboard Autonomous Firefighting Robot (SAFFiR) was designed to be able to move around a ship designed for bipeds and sensors that would allow it to find the heart of a fire it is trying to extinguish and batteries capable of powering it long enough to make its firefighting significant.
The design includes a visible-light camera, gas detector and infrared camera that allow it to see through smoke and find hot-spots.
It's also to be equipped with the ability to throw propelled extinguishing agent technology (PEAT) grenades that are designed to explode near a blaze that is inaccessible to firefighters with hoses, spraying it with fire-retardant chemicals.
Researchers from Virginia Tech and University of Pennsylvania are also writing in algorithms designed to allow the 'bot to work in a coordinated way with human firefighters and follow the orders of human team leaders directing firefighting effort.
That interactive ability should include the ability to respond to gestures and understand signals such as pointing or hand signals.
The prototype will go through initial tests in firefighting mockups on the cashiered destroyer USS Shadwell in September of next year.
Judging by the awkward state of the art in Japanese robots – Honda's gawky Asimo is generally considered among the best of the bipeds – we're still several years away from being able to build anything close to the agility, adeptness and breadth of function the NRL predicts for SAFFiR.
Asimo moves itself around on a demo stage fairly well, but has had trouble even walking up stairs, let alone making its way down a smoky corridor on a ship pitching in rough seas, avoiding injured sailors and obstacles along the way.
This video is of an earlier prototype of the basic bipedal robot frame that will probably become part of the Navy's firefighter.
Considering the most practical semi-autonomous industrial robots are those able to do things like cut the grass on golf-course greens using GPS signals to locate their targets, it will be several years before even advanced engineering projects will be able to deliver autonomous, bipedal robots capable of complex behaviors, the ability to maneuver in a difficult environment and identify the best course of action in situations in which there is more than one choice.
Being able to code in a working interpretation of the kind of sign language humans use during disasters is one thing. That's just pattern recognition and programming.
Being able to identify its own environment, choose an appropriate course of action, propel itself there and take appropriate action under dangerous, ambiguous conditions is the kind of thing the Navy gives medals to sailors for achieving.
Hoping a robot can take over to keep sailors from risking themselves is optimistic and noble, as was the assumption radio-controlled ground-bound robots could help in scouting and cleaning up the Fukushima nuclear plants in Japan.
They made almost no progress at the time, though adaptations and new designs have been more help in the months since then.
A shipboard firefighting robot would be a huge contribution to both the development of effective robotics and a tremendous reducer of the mortal risks faced by sailors.
It is also the kind of leading-edge research project the military is better able to fund (and more easily able to take advantage of) than civilian organizations.
Figuring out how to make SAFFiR work the way it's supposed to would be among the best applications of Navy engineering research money I've heard about recently. It's very likely to succeed, eventually.
It's just not going to happen any time soon.
Now read this:
Carbon nanotubes could let mobile devices use humans for power, just like The Matrix
Why quantum computers will be inaccurate: There is more an one kind of nothingness
Real-world superpowers: Eye surgery lets some see well into the ultraviolet
Virginia Tech./Naval Research Laboratories