Search This Blog

Friday, July 8, 2011

A new generation of animal-like robots is about to emerge from the laboratory


Zoobotics



UNTIL recently, most robots could be thought of as belonging to one of two phyla. The Widgetophora, equipped with claws, grabs and wheels, stuck to the essentials and did not try too hard to look like anything other than machines (think R2-D2). The Anthropoidea, by contrast, did their best to look like their creators—sporting arms with proper hands, legs with real feet, and faces (think C-3PO). The few animal-like robots that fell between these extremes were usually built to resemble pets (Sony’s robot dog, AIBO, for example) and were, in truth, not much more than just amusing toys.
They are toys no longer, though, for it has belatedly dawned on robot engineers that they are missing a trick. The great natural designer, evolution, has come up with solutions to problems that neither the Widgetophora nor the Anthropoidea can manage. Why not copy these proven models, the engineers wondered, rather than trying to outguess 4 billion years of natural selection?
The result has been a flourishing of animal-like robots. It is not just dogs that engineers are copying now, but shrews complete with whiskers, swimming lampreys, grasping octopuses, climbing lizards and burrowing clams. They are even trying to mimic insects, by making robots that take off when they flap their wings. As a consequence, the Widgetophora and the Anthropoidea are being pushed aside. The phylum Zoomorpha is on the march.

Cecilia Laschi and her team at the Sant’Anna School of Advanced Studies in Pisa are a good example of this trend. They lead an international consortium that is building a robotic octopus.
The hug of a monopus
To create their artificial cephalopod they started with the animal’s literal and metaphorical killer app: its flexible, pliable arms. In a vertebrate’s arms, muscles do the moving and bones carry the weight. An octopus arm, though, has no bones, so its muscles must do both jobs. Its advantage is that, besides grasping things tightly, it can also squeeze into nooks and crannies that are inaccessible to vertebrate arms of similar dimensions.
After studying how octopus arms work, Dr Laschi and her team have come up with an artificial version that behaves the same way. Its outer casing is made of silicone and is fitted with pressure sensors so that it knows what it is touching. Inside this casing are cables and springs made of a specially elastic nickel-titanium alloy. The result can wrap itself around an object with a movement that strikingly resembles that of the original.
Arms race
So far, Dr Laschi’s robot is not so much an octopus as a monopus, but she plans to correct this over the next two years by adding seven more arms and a control system that can co-ordinate them all. The aim is to produce an automaton that could help people do difficult jobs underwater—shutting off leaking oil valves, for example.
Another group of engineers at Sant’Anna, led by Paolo Dario and Cesare Stefanini, are also copying an aquatic creature. This one is a fish: a lamprey.
Lampreys are the simplest living vertebrates. Like octopuses, they have no bones (though they do have a rudimentary skeleton made of cartilage). Their nervous systems are simple, too, which makes them a good starting point for studies of the neurological arrangement that eventually spawned the human brain. Sten Grillner’s group at the Karolinska Institute in Stockholm has therefore spent many years studying lampreys, in order to gain insights about vertebrate nerves.
His latest way of doing so is to look at robot versions of the fish. Dr Dario and Dr Stefanini have built him a device called Lampetra (pictured top), which is made of circular segments modelled on the lamprey’s cartilaginous vertebrae. Each segment has an electromagnet attached to it, and these are activated by a current that flows from head to tail in more or less the way that a nerve signal flows in a real animal. A segment therefore first attracts and then releases the next, creating a wavelike movement that propels the robot forward.
Lampetra has eyes, in the form of small cameras, and can use information about colours and shapes that these cameras gather in order to get around obstacles. Indeed, the project’s main aim is to explain how vertebrates use perception to guide their movements. But Lampetra’s unique propulsion system could also have useful applications, as it has proved an efficient way to move a machine through water.
Another zoologist who is using robots to understand the behaviour of real animals is Daniel Germann, of the University of Zurich. He works on clams and is building robot versions of them to find out how the shape of an animal’s shell affects its chances of survival.
Many types of clam escape predators by burrowing into the sea bed using alternating movements of their shell and of the soft, muscular foot contained within that shell. These two parts of the animal take it in turns to act as an anchor, while the other digs further into the sediment. Clams also soften the sediment by expelling jets of water from between the valves of their shells. As a result, they can disappear from view in a few seconds, if threatened.
To understand exactly what is going on, Mr Germann has designed a robot clam. It has a bivalve shell, two strings which move the valves against one another, and a small pump that expels water (he is still working on the best design for the foot). Once he has proved that this arrangement can bury itself in a satisfactory manner, he plans to run a competition between robot clams with different shell shapes in order to understand what makes some more efficient than others. Several of the shells he intends to recreate belong to fossil species. Studying these will test the hypothesis that modern shells are more efficient than ancient ones, and thus out-competed them.
Footfall
Both Dr Grillner’s and Mr Germann’s projects are academically interesting. But other groups are, like Dr Laschi’s, looking for practical applications. StickyBotIII, a robot gecko developed by Mark Cutkosky’s team at Stanford University, is a good example of this.
The ability of real geckos to climb walls and walk across ceilings has long fascinated people. A robot that could do the same would have many uses. The gecko’s secret is that its toes are covered with a forest of fine structures a bit like the ridges of fingerprints, but with deeper chasms between them. Pressed against a surface, the molecules of these ridges attract those of the surface via an electrostatic phenomenon called the van der Waals force. Given a reasonably lightweight animal (or robot) this force is enough to keep the toes attached to the surface.
Like a real gecko, StickyBotIII has four legs. And, like a real gecko, its toes are covered with suitable ridges. The artificial lizard can perform the same tricks as a natural one. Not only is it able to climb a vertical wall, it can even manage one with an overhang.
Nor are clever ways of moving the only use for zoomorphic robots. They are also able to mimic senses beyond the normal human range. Tony Prescott and his team at the University of Sheffield, in England, for example, are attempting to replicate the exquisitely sensitive whiskers of the Etruscan shrew.
These shrews live underground, and rely on their whiskers to feel what is ahead of them when they are moving around in their burrows. By studying slow-motion footage of shrews, Dr Prescott and his colleagues have discovered that the animals constantly sweep their whiskers back and forth, giving anything that seems interesting a second touch.
Using this information, the researchers have built Shrewbot, a robot that reproduces the animal’s head. It has 18 whiskers of different lengths and its software moves these independently of one another, using the information thus gathered to decide whether an object is worth further investigation. So far, Shrewbot can distinguish a smooth surface from a corrugated one. Soon, Dr Prescott hopes, it will be able to recognise basic shapes such as spheres, cubes and cylinders as well. The long-term goal is to build a robot that can operate in places where vision is not much use—smoke-filled buildings, for example.
 Shrewbot within a whisker of a hovering DelFly
All of which is well and good. Robot rescuers would be a useful invention. But the ability that a lot of engineers would really like to put into robots is the flapping-wing flight of insects, with its attendant ability to hover. A tiny flying robot of this sort, equipped with a camera, could get into places that are too small or dangerous for people—enemy bunkers, for example—and report what was going on.
In the Netherlands, researchers at the University of Delft, led by Rick Ruijsink, have developed DelFly, a robotic version of a dragonfly that has two pairs of flapping wings which are moved by an electric motor. DelFly can alternate between flying at high speeds and hovering, in order to take a better look at interesting places. The first version operates mostly under remote control, rather than being a true, autonomous robot. It is, however, fitted with a camera and is able to use information from that camera to adjust its height and direction. The hope is that expanding this ability will lead to true autonomy in the future.
Flights of fancy
Another flying robot, the AirBurr, developed by Jean Cristophe Zufferey at the Ecole Polytechnique in Lausanne, uses a different approach. It does not look much like an insect, but it behaves like one. In particular, it has an insect-like way of dealing with obstacles. Instead of trying hard to avoid them in the first place, it is designed to recover quickly from the occasional thud against the wall, and resume flying. To this end, the AirBurr’s teardrop-shaped wing and small propellers are protected by flexible carbon rods that absorb shocks, and the robot is balanced in such a way that, even if it falls on the ground, it always lands with its propellers facing up, ready to take off again, just like the annoyingly unswattable insect that is buzzing around your room.
The upshot of which is that you do not need to look like an animal in order to behave like one. Robots of the future might end up resembling medieval monsters, with shrews’ heads, octopuses’ arms and lampreys’ bodies. More likely, though, is that specialist machines will be designed to collaborate, with reconnaissance airbots feeding information to groups of groundbots or seabots that are designed to perform different tasks—a robotic ecosystem, you might say.

No comments:

Post a Comment