Acrobatic Octopus Arm Could Be Model for Flexible Robots

原文链接

Though coordinating eight separate arms might seem a tricky task for an octopus brain, what’s really demanding is controlling the arms’ flexible, infinitely variable movements. Now researchers have figured out part of their secret.

Unlike us, specific regions of an octopus’ motor cortex don’t correspond to specific parts of its body. Instead, each region controls different parts at different times. Their motor neural network seems as flexible as their bodies — a phenomenon that expands the range of neurophysiological possibility, and could refine the design of arm-flexing robots.

“We think, because of the complexity of the octopus body and its variability, that it has another way of organizing its control system. That’s what we find in this study,” said Benny Hochner, a Hebrew University of Jerusalem neurobiologist and author of research published Thursday in Current Biology.

“It’s suited to a structure with many more degrees of freedom than our own body, which is constructed around a segmented skeletal structure with few degrees of freedom.”

How octopuses control their arms has been a focus of Hochner’s work for more than a decade. In earlier studies, he helped show that seemingly complex movements are actually combinations of individually simple motions. Hochner also found that many of the movements are guided peripherally, rather than by the brain, as if each arm had its own spinal cord.

An octopus brain sends a general prompt, and the arm computes the specifics: It’s much simpler than running all those calculations in the brain itself. And all this is especially interesting to roboticists who want to build machines with flexible appendages, ideal for rescue bots working in disaster areas or surgical machines weaving through a body.

“The idea is to draw inspiration from biology to answer the question of how to generate movement in a flexible structure, and how to control this with the nervous system,” said Hochner.

In the latest study Hochner’s team ran electrical currents through wires inserted into in the animals’ brains, measured the resulting movements, and then dissected the sacrificed animals to see exactly what the electrodes had stimulated.

They found yet another example of modular, highly efficient design: Each site proved capable of generating different movements, in different arms, with movements becoming more complex as the current increased. In humans, most body parts are controlled at a single, unchanging location.

“The networks are embedded in one another. The system is remodeled according to stimulation. It’s more dynamic, rather than strictly organized,” said Hochner.

Hochner suspects that other neurological programs, stored elsewhere in the octopuses’ bodies — perhaps at the base of each arm — act as gates, blocking signals from the brain or allowing them to pass.

That possibility is especially intriguing to Cecilia Laschi, a biomedical engineer at Italy’s Sant’Anna School of Advanced Studies and member of the Octopus Project, a group of researchers building octopus-inspired soft-bodied robots.

“This is very important for robotics. If you build a robot with many degrees of freedom, it becomes very difficult to control.” said Laschi, who was not involved in the study. “We know that some movements are controlled peripherally, some parameters are set by the brain, and we will do the same thing in our robots.”

But whereas roboticists building humanoid forms can already try to mimic the human brain’s layout in their computing, Laschi said that “with the octopus, we’re not at that level — yet.”

Citation: “Nonsomatotopic Organization of the Higher Motor Centers in Octopus.” By Letitzia Zullo, German Sumbre, Claudio Agnisola, Tamar Flash and Binyamin Hochner. Current Biology, Volume 19 Issue 18, September 17, 2009.

Secret Law of Flying Could Inspire Better Robots

原文链接

A unifying theory of winged locomotion could explain the magical mid-air maneuvers of birds and insects, and guide the design of flying robots.

Using high-speed video, biologists modeled how hummingbirds and hawkmoths use asymmetrical flapping to make slow, mid-air turns. The model predicted how five other flyers turned at full speed, hinting at a universal turning technique for flying creatures. 

"It’s basically an exponential damping system," said Ty Hedrick, a University of North Carolina animal aerodynamics expert. "The strength of braking increases in proportion to speed."

Though scientists understand the principles underlying many flight-enhancing physiologies, from birds’ hollow bones to dragonflies’ flexible wings, the biomechanics of turning was in many ways a mystery.

Researchers were unsure whether different species used fundamentally different mechanisms, or variations on a basic theme. Hedrick’s findings, published Thursday in Science, describe a common solution shaped by evolutionary pressures in the 150 million years since dinosaurs took to the air.

Though the dynamics probably can’t work at large scales — building-sized robotic birds won’t ever be as agile as a swallow — they could be harnessed in small drones used by explorers or the military. Compared to the average hummingbird or fruit fly, such craft are now clumsy and unstable. 

"The results will inform all future research into maneuvering flight in animals and biomimetic flying robots," wrote University of Montana, Missoula biomechanicist Bret Tobalske in an accompanying commentary.

Hedrick’s team used 1,000 frame-per-second video cameras to watch hawkmoths and hummingbirds hovering before a feeder. As each turned away, one wing flapped faster on its down-stroke, while the other flapped faster on its up-stroke.

The asymmetry causes flyers to lose speed as soon as they start to turn. The effect is strongest when velocity is highest.

"The moment they start turning their wings and stop symmetrically flapping, their bodies act like a brake," said Hedrick.

Measurements of the motion provided a model that, adjusted for size differences, predicted the mid-air turn motions of four insect species, a cockatoo, a hummingbird and a bat.

In animals with proportionally similar bodies, rates of wing flapping — not body size — controlled turning ability. Agile hummingbirds and fruit flies flap their wings the same number of times to complete a turn.

"To understand the importance of this result, consider the array of solutions that flying animals have at their disposal to modulate aerodynamic forces," wrote Tobalske. "The fact that the flapping counter-torque model is robust over a wide range of body size indicates that it represents a universal model," he wrote.

The effect probably helps flyers regain equilibrium when hit by gusts of wind, providing a natural stabilizer that engages before their brains can react to a perturbance, said Hedrick.

The study’s other co-authors, Darpa-funded University of Delaware mechanical engineers Xin-Yan Deng and Bo Cheng, will use the findings to refine their insect-inspired unmanned aerial vehicles.

As for Hedrick, he next plans to study mechanisms used in more complicated aerial maneuvers, perhaps equipping swallows and other small birds with sensor-filled backbacks.

"Animals are doing things so smoothly and gracefully that we don’t even realize that these are very hard tasks," Hedrick said. "In a robot, we have trouble replicating that behavior."

Citations: "Wingbeat
Time and the Scaling of Passive Rotational Damping in Flapping Flight."
By Tyson L. Hedrick, Bo Cheng, Xinyan Deng. Science, Vol. 324, April
10, 2009.

"Symmetry in Turns." By Bret W. Tobalske. Science, Vol. 324, April 10, 2009.

The next big fin East meets west Rise of the robogeeks Why do people die that way? Clever fools: Why a high IQ doesn't mean you're smart