40°12'21.6"N 2°00'38.5"W

Before the drone could take flight, scientists first needed to pluck the feathers of a pigeon.

Human-built flying machines have, since the dawn of powered flight, relied on fixed wings or spinning blades to stay aloft. When animals evolved flight, they did so without dedicated flaps and ailerons, and they especially did it without light-weight carbon fiber frames.

Real feathers filling out the wing of a drone is a way to explore the benefits of bird-wings in powered flight. The drone, built by researchers at Stanford University, is first and foremost a scientific exploration. If the lightness and the silence of bird feathers can be modeled, they could perhaps be reverse engineered. If those wings can be reverse engineered, then maybe ornithologists could do science with ornithopters.

And maybe, just maybe, the military could pilot feathered drones as spies in disguise into the battlefields of the future.

I’m Kelsey D. Atherton, reporting from Albuquerque, New Mexico, and this fortnight I want to talk about fighting battles in the Uncanny Valley.

“Uncanny Valley” is a term that comes from Japanese roboticist Masahiro Mori in 1970. It is the feeling of unease when artificial beings approximate the feeling of something alive, but not quite. The valley is the gulf between symbolic cartoonish representations and realistic depictions. It is the not-quite-there state. It is the feeling of bad computer-generated graphics from the early 2000s, or of semi-animate dolls, or perhaps most relevantly, of watching a legged robot stumble forward with the mass of a large dog and the flailing mechanics of a spindly insect.

It is also a near-inevitable result of all forms of biomimicry in robots.

Some animal-shaped robots, like Tomorrow Wars reader favorite the Russian Owl Drone, are extremely goofy. Others, like the swimming MantaDroid, pose a more realistic chance of being confused for the real deal. All carry with them the danger that they trade whatever utility initially gained in combat for a retaliatory scorched-earth approach, where jumpy soldiers fire on all animals on the off-chance that the birds are in fact secretly robots.

More even than the harms to wildlife, robots traversing the uncanny valley warrant skepticism because of the limited utility of bending biological forms to war. Whatever gain in silence or efficiency the military might gain from a pigeon-feathered flying machine could be lost to the robot catching the eye of an eagle, which thrashes the robot mid-flight.

There is also a danger in hewing too closely to designs modeled after what nature has already done. The prologue to the history of powered flight is filled with sketches and contraptions and briefly airborne corpses, all dedicated to the idea that flapping and feathers, not thrust and lift, were all that stood between people and the sky.

Instead of imitation, biology can inform wholly new mechanics. These machines won’t leave the uncanny valley, but they will navigate it in an entirely different unsettling way.


If the feathers of a pigeon drone aren’t enough weirdness, consider the flapless wings of an Air Force Research Laboratory design. Approximating the skin-like membranes used by flying mammals, the flapless wing skips the rough edges of modern design and instead manipulates its shape by pushing against the smooth, plastic covering. While unmistakably a plane, the wings push and flex and warp, as if made by a horror director who decided to dabble in aeronautics.

For all the weirdness, the design promises a fuel savings of 10 percent, thanks largely to reduced drag.


The battlefield is a bad place to break a leg. For a legged robot, a broken leg could mean abandonment or destruction, and a troop of humans who now have to carry everything useful previously off-loaded to the robot. A new kind of healing joint allows legged robots to break specific joints, and then reforge those connections to between 30 and 80 the strength of the unbroken joint. As demonstrated, the healing took place in 30 minutes, likely far faster than repair under fire, and without and without any need for direct human intervention. It is this kind of biomimicry, of aping capabilities more than form, that is perhaps the most exciting.


For all that is interesting in machines that resembles animals, the most likely developments in military robots are machines that resemble existing people vehicles. Russia’s Shturm and Soratnik robot tanks suggest a future of armed automation that most closely resembles the recent past. Removing people from inside the vehicles opens possibilities for commanders, provided they are willing to trust the machines to follow-through.


That’s all for this week, my Tomorrow Warthogs. Any tips about robots that look like animals, or animals that look like robots, email me at katherton@c4isrnet.com.