Skip to main content

Biomimetics: taking cues from nature

How can a butterfly’s wings help drive down the cost of transmitting information across the Internet? Or a burdock burr lead to a revolutionary new fastener? As scientists and engineers around the world have known for some time: the natural world is still the best source for inspiration.

[This article is an extract from "Innovations #3"]

The iridescent wings of the butterfly Morpho Rhetenor have puzzled scientists for years. How do they achieve such an intense blue colour? Master painters such as Giotto had to crush the precious stone lapis-lazuli to obtain the pigment. But the butterflies can’t crush stones to cover themselves. So how do they do it?

Under an electron microscope, scientists in the Thin Film Photonics Group at the University of Exeter discovered that the butterfly’s wings,reveal tiny comb-like structures, about the same size as the wavelength of light itself. This allows them to interact with light very strongly indeed ,and gives them their iridescent blue colouring.

A similar iridescence was found on some marine worms and has been traced to a series of microscopic holes in the creature’s body hair.

Now, a similar miniature mesh has been created artificially by Alfredo de Rossi of Thales Research & Technology Physics group, to control light in a very specific and potentially very useful way: “We are not the first to use light to control light, but we are attempting to do it in a smarter way,” he explains.

The mesh confines light to a tiny volume, producing an extremely high density of light energy. The concentration generates an electrical field so strong that it can then be used to control other beams of light.

“In a transistor, you have three terminals. The current on one terminal is used to control the flow of current across the other two,” de Rossi explains. “We have found an approach to do this with light. We don’t use electricity except to power up the equipment. All signals are carried on light.”

Removing much of the need for electricity gives this technology the power to transform communications, which is currently expanding at an unsustainable pace. “Communication technology is consuming more energy due to the dramatic rise in communication traffic.

According to many who study power consumption statistics in the telecoms domain, if this carries on, we will soon have to use all the electrical power that we produce just powering the Internet. Clearly this is impossible,” says de Rossi. The new systems that he and colleagues are pursuing use just milliwatts of power, offering enormous potential for the future of our increasingly interconnected world.

The eyes have it

This “biomimetic” approach to solving a pressing problem, by emulating aspects and systems from nature, seems to be inspiring a raft of new projects around the world and producing remarkable results.

Researcher Jean-François Goudou and his team at Thales are two years into a project aimed at recreating human vision. It may appear that cameras, with their lens and detectors, already offer an acceptable imitation but this is not so, says Goudou. The eye is far more complicated.

“The retina is not only a photon collector, it also processes the data. It does not provide an image to the brain but transmits information about the spatial and temporal features of what you are seeing,” says Goudou.

The retina’s processing includes de-noising, contour recognition and orientation, as well as movement recognition. This is much more useful to a data-processing device such as the brain (or a computer) than an image which must first be “scanned” for its details before it can be processed.

Another difference is that the pixels of a camera see the same field of view at the same resolution, whereas the eye has a very high resolution area in the centre, but less detail on the periphery.

Goudou believes that a successful biomimetic system may be another few years away but, when it comes, it will make computer vision a much more efficient thing. It has obvious applications in robotics, which need to understand their surroundings quickly to be as responsive as possible.

“This would allow a robot’s eyes to move very quickly to objects of interest in their field of view,” says Goudou. “The robots would also receive motion information about things in their environment.”

This would be a step towards the ultimate biomimetic goal of producing artificial animals. Consider the exploration of other worlds: at the moment, rovers trundle across distant planets, sending images of their surroundings back to “drivers” on Earth, who help them navigate the alien landscape.

This is a long-winded process and the use of wheels severely restricts where the rovers can go. Even the most modern rover, such as NASA’s Mars Curiosity, is terrible at coping with unexpected situations. What are the alternatives?

“Talk to any biologist and they will tell you that the most impressive and adaptive systems are humans and animals. To survive, animals are capable of displaying astonishing adaptive capabilities,” says Christophe Meyer, senior expert, director of research and advanced studies, secure communication and information systems at Thales.

What if you could design your rover as an artificial animal – something referred to as an “animat” – that can walk or crawl across any surface, recognising hazards and avoiding them? These could also be used on Earth in situations that are too dangerous for humans.

Walk this way

Meyer’s interest in biomimetics began early. His father worked with the pioneering MIT researcher Rodney Brooks, who set about developing adaptive robots. Meyer remembers one particular robot that was engineered to learn how to walk: “This robot built its own programme, allowing it to learn how to cover the largest area in the smallest amount of time,” says Meyer. The software was designed to learn how best to achieve its objective, rather than just slavishly following a set of commands – when one of the robots had a leg removed to simulate an accident, it relearned how to move. It may not have been as efficient, but it learned how to make the best of what it now had and was able to continue performing its mission. As well as the hardware biomimicry – for example, legs rather than wheels – there must also be biomimicry in the software. This leads to a new approach in the field of artificial intelligence.

“In the old days, it was all about beating the Turing test,” says Meyer. This was a concept introduced by Alan Turing in 1950 in a paper entitled Computing Machinery and Intelligence. He posed the question, “Can machines think?” and proposed a test by which a person holds two conversations via computer screen and keyboard with participants that cannot be seen. One participant is a computer, one is a human. The computer passes the Turing test if the human asking the questions cannot determine which of the participants is human.

 

Modern artificial intelligence does away with this lofty goal and simply concentrates on giving the machine enough smarts to achieve a goal. It doesn’t tell the machine how to do so, instead it lets the machine work out the details for itself based on what it can sense of its environment. In other words, it makes it up as it goes along.

Meyer is in charge of Thales’s adaptive systems and biomimetics simulation project, which uses biomimetic software to simulate the behaviour of human beings in virtual environments. The SE-Star software allows building designs to be tested for their efficiency, safely and user friendliness before they are built.

The project began five years ago when Meyer realised that human irrationality and behaviour in strange situations has been studied sufficiently well that it could be simulated on a computer.

“We can now use the system to test critical infrastructure design before building the real thing. Then we populate this virtual environment with virtual people who behave realistically and see what happens,” he says. “For example, we can put a fire or smoke anywhere we want in the environment and see how people behave.”

They can then change the placement of the exits and run the simulation again to see if more people can get out quickly.

The same software tools can also be used for other applications.

The first is to train people who use computer screens to monitor the movement of people, such as CCTV operators of crowd management systems.

Currently, operators are trained by sitting with colleagues and watching daily operations but, to build expertise, they need to learn how to handle difficult situations. Simulations of real human behaviour offer a clear advantage.

“We call this embedded training because you use your operational system but instead of being connected to reality, you are connected to a biomimetic simulation,” says Meyer.

It is rather like pilots in flight simulators who practise emergency procedures, even though most will never encounter such a situation in real life.

Another application in development is decision support. Simulation software can be used to test possible solutions before they are implemented.

It works by taking a snapshot of a real-world situation – say the number of people in an unexpectedly busy airport – and transferring this to the simulation, where possible scenarios can be tested quickly.

For example, additional X-ray machines could be opened or closed, staff could be moved from one operation to another.

“We do not speak about automatic decision making because the system cannot say what will happen, only what could happen,” says Meyer. Nevertheless, it could help operators make good real-time decisions.

“Taking inspiration from the microscopic world, artificial immune systems are now developed in my team by Fabien Flacher to provide new adaptive cyber security functions, able to detect complex intrusions in critical information systems as well as to dynamically follow the evolution of these systems,” adds Meyer.

Biomimetics is a broad, sweeping discipline – from the smallest component to the largest system, answers can be found in nature. All engineers and technologists need to do is look for them.

After all, nature has had four billion years of evolution to solve all sorts of problems.

“This is just Darwinism speeded up,” says Meyer. As the old saying goes, imitation is the sincerest form of flattery – and Mother Nature should feel very flattered indeed.

The heart of biomimetics
Biomimetics transforms nature’s solutions into technology. Although its origins can be traced to Leonardo da Vinci’s drawings of bird-like wings designed to allow humans to fly, it was transformed into an academic field during the 1950s by American scientist Otto Schmitt. He coined the term as part of his doctorate, in which he designed an electrical circuit known as a “Schmitt trigger”, which uses feedback to convert an analogue electrical input signal into digital output. Schmitt was inspired by the way a neural impulse moves through squid nerves.
The Schmitt trigger is not the only biomimetic success story. Perhaps the best known is that of Swiss electrical engineer George de Mestral, whose dog became covered in burdock burrs during a hunting trip in the Alps in 1941. He examined a burr under a microscope to discover that its grip is simply the product of tiny hooks. With this as his inspiration, de Mestral invented a unique fabric hook and loop fastener – better known as Velcro.