Perception is one of the major research fields in the cognitive sciences. Our understanding of the world is largely based upon the sensory information we gather and the information processing done by the brain. Some sensory modalities are studied in much detail like vision, other modalities are still less understood.
In rodents like mice or rats, which move a lot in faint light or darkness, an important sensory input comes from their whiskers. Mice use their whiskers to actively explore their environment. For example, they are able to perform discrimination of surfaces by only using their whiskers, or to follow walls by keeping contact to the wall with their whiskers.
Mice and rats have their whiskers arranged in rows and arcs on both sides of their face. Each whisker is surrounded by spindle-like touch receptors. While the exact properties of the sensory signals are not known in detail, much research has been done on the cortical representation. Each whisker is represented in a "barrel", a cortical area that reaches through several cortical layers. These barrels are topologically arranged, each barrel corresponding to one whisker on the contralateral side.
Figure: The upper picture shows a rat's snout with its whiskers. On the bottom panel one can see the cortical barrel representation.
Whiskers as a Model System for Biology
Whiskers are superb tactile organs. The tactile resolution that rats have with their whiskers is almost as good as our fingertips, which is amazing considering that we have a very high spatial density of tactile receptors in the fingertips, whereas rats have only about 30 whiskers on each side of their face. Whiskers have been a popular model system for tactile processing in biology for a long time for several reasons: rats are popular laboratory animals (quite intelligent, easy to breed and handle, cheap), but more important: stimulating a whisker is very well defined. Stimulating skin is much more difficult to reproduce. Imagine stimulation of your fingertip. If it has to be repeated, it is quite difficult to stimulate the exact location again, while the whiskers are distinct from each other, they are numbered in columns and rows making experiments well reproducible. Furthermore, each whisker has an anatomical region that receives its input. This is also well-defined and helps to correlate sensory stimulation to neurophysiological recordings.
Whiskers in Robotics
Despite this interest from the biological side, in robotics whiskers have been neglected just until about 3 years ago. But tactile sensors are of great interest to roboticits as well: they are independent of light conditions, so complementing visual camera input. They give detailed information (textures, shapes) of the immediate surroundings and they have potential for commercial applications where light-independent sensors are needed, e.g. for survey of pipes.
While there are a multitude of cameras and microphones with which one can technically mimick the visual and auditory sensors, there are no sophisticated whiskers available. The goal of the Artificial Mouse Project is to develop an active whisker system mounted on a moving robot.
