Machines That Mimic Nature
Re-inventing ideas from nature is nothing new; think sonar in ships and radar in aircraft, both of which are analogous to echolocation in marine mammals and bats. Indeed, borrowing ideas from nature dates back centuries; for instance, see our blog post “The TENS Commandments” that describes how ancient Greeks and Romans advocated harnessing electric shocks from the Torpedo ray as a system of pain relief.
Since then humans have frequently tried to emulate nature with varying outcomes: examples include Leonardo da Vinci’s less-than-successful exploration of flying machines with flapping wings. In contrast, the modern phenomenon that is Velcro, invented in 1948, was inspired by the way that burdock burrs kept sticking to the clothes and dog’s fur of inventor George de Mestral while he was out hunting.
One invention that da Vinci did successfully see through to fruition was his so-called robotic knight, an early foray into the world of robotics. Renaissance genius, da Vinci was fascinated by human anatomy; he spent long hours dissecting corpses in order to discover how the human body worked. This gave him an understanding of the human skeleton works and the way muscles interface with bones. Applying these same principles to a machine, he actually built the robotic knight, primarily as entertainment for parties thrown by his wealthy patron Lodovico Sforza.
Although this robot – which was driven by a system of pulleys and gears – has not survived, it was reputedly capable of walking, sitting down and working its jaw. Modern-day robotics expert Mark Rosheim has used da Vinci's notes to build a working model of the contraption. Bringing the concept right up to date, Rosheim subsequently used some of the original concepts to design planetary exploration robots destined to be used by NASA.
Biomimicry is now big business, but is proving an especially fruitful field of exploration in relation to robotics. Here are five concepts that are being transferred from the natural world to the world of machines.
- Robotic fish
Two features of the ghost knifefish could prove useful for underwater exploration and search and rescue: their ability to sense their environment through perturbations in an electric field and the way that they move through tangled underwater “forests” using their long undulating fin. Electric sensors scattered across their body mean these fish can detect things coming from all directions. At the same time, the ghost knifefish can swim forwards or backwards by changing the direction of undulations on its ventral fin, and it also has the ability to use counter-propagating waves that meet in the middle of this fin to move upwards.
Now, researchers from Northwestern University in the United States are demonstrating artificial sensory and locomotion capabilities on two separate robotic platforms. The aim now is to bring them together into a single working device.
- Buzzing about like a mosquito
As arguably the most successful flying animals, insects have been studied intensely in our quest to develop aerial machines. One development is the Gimball – a drone designed to deal with crashes, and to right itself after a collision. This flying robot is able to bounce off walls and tree trunks thanks to its protective spherical roll-cage and gimbal mountings, and it moves in a similar way to a mosquito. A team at Switzerland’s Ecole Polytechnique Federerale de Lausanne has developed the machine for use in disaster situations, such as entering a building on fire, or after a radiation leak. Eventually, it will also incorporate artificial intelligence to enable it to operate autonomously.
Meanwhile, a team from Harvard in the US has created an agile fly-sized robot. The carbon fibre Robo-fly weighs a fraction of a gram and has super-fast electric "muscles" to power its wings. Again, the hope is that they will eventually be involved in rescue operations.
- Overcoming turbulence
Speaking of insects in adverse situations, how do bumble bees keep flying in windy conditions? One longstanding urban myth about is that they shouldn’t technically be able to fly (see our blog “Flight of the Bumblebee”) but in fact they are all-weather foragers and so have to be excellent flyers. So good, in fact, that scientists decided to film them in a wind tunnel to learn more about how to maintain stability for micro air vehicles operating in bad weather. The researchers found that the bees reduce their speed when there’s turbulence in order to spend more energy making flight corrections, and they are particularly vulnerable to side winds. Further research will enable scientists to apply these lessons to the design and control systems of future drones.
- Insect skyscrapers
Termites are well know for building impressive structures that incorporate natural “air-conditioning” systems as part of their nests. These insects are able to build “cooling towers”, which are several metres high, even though they can only follow simple rules and have no architect to guide them.
A Harvard team has developed robotic “bricklayers” capable of building large, coherent structures simply by sensing their immediate environment and taking limited cues from one other. This decentralised approach to robot programming offers some major advantages over very sophisticated systems, the researchers say. For instance, a swarm of such robotic bricklayers could operate in a hazardous environment such as a flood zone building sandbag dams or eventually in space. The advantage is that, even if one or more of these relatively simple construction bots are damaged, the remainder can continue to operate as a team to complete a task.
How do they do this? One approach is to take a leaf out of the termites’ book where work is coordinated via chemical scents or pheromones as well as the actual shape of their developing construction projects. This is a process called “stigmergy”.
- Keeping in step
Of course, insects are not the only invertebrates to provide us with robotic inspiration. Studies of the neural networks controlling the legs of invertebrates like crabs and lobsters have revealed the rhythmic nerve impulses that underpin this movement. These rhythmic impulses are called central pattern generators or CPGs and are among the best known of all neural circuits.
Building their electronic equivalent enables us to control movement in robots in new and exciting ways, for instance by mimicking the movement of animals with lots of legs. Such multi-legged animals are especially popular models in robotics because they are very stable. What’s more, the repetitive motion effectively runs on “autopilot” once started and the control systems are relatively simple. The CPGs are also modular and can be linked together to create more complex behaviour.
Where will Nature inspire Scientists next? We'd love to hear your thoughts on this fascinating topic in the comments below.