Why the praying mantis may hold the key to teaching robots how to see
Scientists discover the insect’s unique way of perceiving depth could solve computing problem
ACCURATELY judging distance is one of the biggest challenges facing modern robotics.
Now, however, British scientists claim to have made a breakthrough thanks to an unlikely role model: a praying mantis wearing red glasses.
Currently, robot “vision” systems, such as those used by drones to pick up packages or navigate around objects, tend to mimic human stereo vision.
Each eye sees a marginally different view of the world and the brain merges the two views to create a single image, using the differences between the images to calculate how far away objects are.
While it works well for humans, artificial systems based on the same concept require significant computing power, which both slows and weighs down robots.
Researchers at Newcastle University therefore looked for other examples of stereo vision in the animal kingdom and focused on the praying mantis, the only insect known to possess it.
To investigate the bug’s sight system, the team created bespoke 3D red glasses which they temporarily glued on with beeswax. They showed the insect moving images of prey, as well as complex dot patterns which are also used to investigate human 3D vision.
The experiments found that, because mantises only attack moving prey, their neurological processes do not bother to compare the details of still pictures in each eye. Instead they judge distance by simply looking for places where the picture is changing, and they do so better than humans.
“This is a completely new form of 3D vision as it is based on change over time instead of static images,” said Dr Vivek Nityananda, a behavioural ecologist. “In mantises it is probably designed to answer the question ‘is there prey at the right distance for me to catch?’.”
Published in the journal Current Biolog y, the study concluded that the praying mantis vision system was “very robust” and simpler than that used by humans, meaning it could provide a far better template for robots.
Dr Ghaith Tarawneh, who worked on the research, said: “Since insect brains are so tiny, their form of stereo vision can’t require much computer processing. This means it could find useful applications in low-power autonomous robots.”
A significant proportion of robotic research is dedicated towards drone technology, where simplicity and lightness of operating systems is at a premium.
In December 2016 Amazon successfully trialled its 30-minute fully-autonomous drone package delivery system for the first time, a model that is thought will lead to the dramatic increase of the number of robots in the sky.