The rat’s whiskers

In designing intelligent machines, research on sensing technology has focused most on vision, then hearing, then, perhaps, chemosensation (smell and taste).

Although touch sensors are employed on many robots, their role has usually been to support simple, if important, functions, such as detecting unexpected contacts so as to prevent robots from damaging themselves or others.

This stands in interesting contrast to the use of tactile sensing in the animal kingdom, be it the human fingertip or the sensitive tactile hairs found on many animals. In nature, touch is used, not just as an alerting stimulus, but also to solve complex perceptual tasks: determining the shape, texture and position of encountered objects; deciding whether something is moving and, if so, how fast and in what direction; distinguishing soft from hard, hot from cold, living from non-living.

The FP7 Biotact (biomimetic technology for vibrissal active touch) project aims to change the way that tactile sensing is thought of by the robotics community, and to demonstrate that biologically inspired tactile sensors can be used to put intelligent machines “in touch” with the world as never before.

Most people think of the skin, and particular of fingertips and lips, as the supreme organs for tactile sensation; however, in the natural world, many animals do touch sensing at a slight distance, using long hairs, or vibrissae, to explore their surroundings.

A vibrissal sensor works rather like an old-fashioned record needle - the bumps and troughs of a contacted surface are translated into movements of the vibrissal shaft, and these, in turn, are detected by hundreds of pressure-sensitive receptors inside a specialised hair follicle.

One of the benefits of this arrangement is that, unlike a fingertip, the delicate sensory transducers (the receptors) are kept away from the contacted surface where they might otherwise sustain damage from the direct physical contact needed for touch sensing. This is an attribute that could be useful in artificial tactile systems where wear-and-tear of the sensing apparatus is a significant problem.

The part of a vibrissa that wears out is its tip, which in animals and potentially in robots, can be continuously replaced through new growth.

Animals that specialise in the use of vibrissal sensing include rats and mice, seals and walruses and some of the smallest living mammals, the shrews. In some of these species, particularly those that are nocturnal or live underground, the facial vibrissae, or whiskers, are a more important sense organ that the eyes.

Rats can sense texture using their whiskers with similar accuracy to the human fingertip, and the whiskers of the shrew allow these animals to detect, recognise, track and catch prey insects with lightning speed.

It is the prospect of putting these kinds of sophisticated tactile sensing capabilities on to robots that excites researchers on the Biotact project. The consortium, which is made up of researchers from nine groups in seven countries, includes biologists studying vibrissal sensing and vibrissal neural processing in whiskered animals; computational modellers developing biologically inspired algorithms for extracting meaning from vibrissal signals; and roboticists, designing, building and testing novel artificial sensors that use arrays of vibrissa-like fibres.

The roboticists are building two prototype systems to investigate the potential for such devices. The first is an array of artificial vibrissae, termed a Biotact sensor, which will be attached to the end of a robot arm and will be tested in tasks with potential industrial relevance, like sorting objects according to their texture.

The second is a rat-like robot that will have an agile head and a wheeled base and will use its whiskers to track and follow small fast-moving targets.

This machine will provide a strong test of whether we can build robots that can match the speed, agility, and discriminative capabilities of whiskered animals. One of the most striking characteristics of the vibrissal systems of rats and shrews is that their whiskers are not passive sensors waiting to be deflected by an encounter with an object.

Rather, these animals continuously and actively move their whiskers back and forth at high speeds in a behaviour known as whisking. Since whisking requires energy, it must have some important benefits.

Biotact researchers believe that the ability to position the whiskers accurately, and to employ alternative whisking strategies in different contexts, may constitute the principal gains.

In other words, rats may whisk for the same reason that people repeatedly adjust the position of their fingertips when exploring objects with their hands - because we get the best sensory information when we can carefully control how our sensors interact with the world.

An artificial vibrissal system that is actively controlled in this manner will be very different from the blundering binary collision detectors that provide the tactile sensing competence of most contemporary robots. The Biotact sensor will, we hope, feel its way, dancing around a target object with purpose and accuracy to extract a rich tactile percept of its contours.