Researchers at The Vision Centre and Australian National University have achieved a world-first in working out how fiddler crabs perceive their world and respond to it.
Their research is also expected to assist in the design of better machine vision for robots.
Fiddler crabs have life-and-death decisions to make, says Dr Jan Hemmi. They must instantly detect a predator, a mate, a competitor or landmark in their flat, beach world – and act accordingly. “Our work is aimed at understanding how they process the visual signals they receive and convert them into behaviour – a process common to all visual organisms, ourselves included.”
Fiddler crabs have virtually all-round vision, including overhead, provided by 9000 separate eye facets, or ommatidia. Working with Dr Hemmi, researcher Jochen Smolka is the first person to map the crab’s visual field, work out how its different parts dictate the animal’s behaviour and help it interpret what it sees.
“Unlike our eyes, the crab’s eyes do not move, so it uses different parts of its visual field for different tasks. Some require a sharp focus and some of which require just a general lookout to be maintained,” Dr Hemmi explains.
“In the fiddler crab, Jochen has demonstrated that the sharpest vision is in the horizontal plane immediately in front of the animal. Its eyes here are really adapted to fine detail. This it uses for identifying and communicating with potential mates.”
The crab also sees quite well horizontally to either side with especially good spatial perception enabling it to see how far objects are from one another. It uses this to keep watch for rival crabs and monitor how far it is from its own burrow, so it can run for cover.
The eye cells that make up its overhead and rear vision are much more thinly spread, sufficient just to provide warning of the approach of a predator like a gull, outlined against the bright sky. These provide no fine detail, only bright/dark signals. “The crab only needs to see one dark spot moving in its upward vision to know it must run for its burrow,” Jan says.
To test the response of crabs to objects glimpsed above them, the team has also constructed a ‘crab treadmill’ – a ball suspended on a column of air that tracks the direction the crab scuttles when it sees something scary, using the same principle as a computer mouse.
Crabs also see their world in unusual colours, they say. The beach where they live is drenched in ultraviolet light, and the team has found that crabs can see in the near ultra-violet as well as other colours. This ability may provide a way to recognise mates or rivals based on UV patterns on their shell, they say.
“All vision has tradeoffs including our own,” Jan says. “There are areas where we need very sharp vision for detail, and areas where we need much less acute vision to warn us of danger. No creature has acute vision all round.
“In the case of humans, all our acute vision is provided by the fovea, a small pit at the back of the eye where vision cells are thickly concentrated, enabling us to read, recognise faces and see detail. In the case of the crab, it has a concentration of vision cells a few degrees above and below the horizon and towards its front, where its most acute vision exists.”
“So while the crab may seem a simple creature, its vision is exquisitely adapted to the featureless mudflats it inhabits, providing it with all the information it needs to navigate, feed, mate, fight rivals and dodge predators.”
The crab’s eyes may hold important lessons for robot design, Jan says. “Most robots use TV cameras which are like the human eye and provide a flood of information which is hard to process quickly. The crab’s eye, on the other hand, performs all the essential tasks needed in autonomy, but with far less information being processed.”
“For certain types of robot, this type of machine vision may be far more practical and appropriate – and the crabs can teach us much,” he says. The team plans to build a panoramic sensor based on the crab’s eye to test in robots.
Part of the program at The Vision Centre in which Dr Hemmi’s team works is devoted to applying the principles used by creatures such as honey bees and crabs in their vision to robotic aircraft and other autonomous devices.
Editor's Note: Original news release can be found here.