金沙官网网址大全

'Sniffer-bot' algorithm helps robots seek scents

作者:樊簖李    发布时间:2019-02-27 02:19:02    

By Mason Inman Moths are renowned for their ability to pick up a faint whiff of pheromones from faraway mates. Robots may soon match this feat with the help of a new mathematical method developed to help guide them toward a scent. In tests, the algorithm made virtual scent-hunter bots move just like moths do, snaking and spiralling toward their goal. Finding the source of a strong smell is pretty straightforward. As you move, simply check whether a scent is getting stronger or weaker and head in the direction in which it is strongest. But the situation is much harder when the scent is just barely detectable. If there are only a few intermittent hints of a scent, then there is no clear trail to follow. Add air turbulence, and the problem gets even harder, since eddies can break up the scent, creating odd-shaped pockets of detectable scent and gaps containing no odour at all. Now a group of physicists have devised a relatively simple mathematical method for guiding a scent-tracking robot that can cope with these difficult situations. Massimo Vergassola at the Pasteur Institute in Paris, France, and colleagues created an algorithm that tells a robot how to move in order to gather as much olfactory information as possible. This allows it to home in on even the faintest of scents. The researchers tested the algorithm using computer simulations. They used both real and simulated data on scents floating through turbulent air, and let a virtual robot loose to try to find the source of the smell. “When the first patterns [of movement] came out, I was pretty amazed,” Vergassola told New Scientist. Despite the simple rules built into the algorithm, the virtual robot moved in complex ways reminiscent of moths or birds following a smell. Just like these animals, the virtual robot would sweep back and forth in S-shaped curves and perform spirals as it gradually homed in on the source. The algorithm uses more than just information about when a scent is picked up. It also takes account of the times when it the smell is not picked up, which shows the directions to avoid. This helps it balance between two extremes – heading straight toward where it guesses the smell is coming from, and making wandering movements that help it collect information but do not take it closer to the source. The algorithm would be straightforward to run on a real robot, Vergassola says. It could also be used for other tasks that involve searching with limited information, for example finding the ideal branches to use when sending data through a network. Alan Gelperin, at the University of Pennsylvania in Philadelphia agrees, saying that the study “provides a new framework for understanding a large and significant class of problems encountered in real world situations”. Gelperin also agrees that it should be feasible to run the new algorithm on a robot outfitted with chemical smell sensors. Adding instruments to detect the direction of air flow around the robot could make the method even better, he adds. Journal reference: Nature (vol 445,

 

Copyright © 网站地图