‘Sniffer-Bot’ Algorithm Helps Robot Seek Scents
Researchers in France have developed an algorithm that could allow robots to find the source of a faint scent even in the midst of air turbulence, much like a moth does.
Massimo Vergassola and some of his colleagues at the Pasteur Institute in Paris tested their simple algorithm in a virtual environment and found that it not only allowed a virtual robot to successfully track and find the source of a scent, but it caused the virtual robot to move in complex back and forth sweeping motions, s-curves, and spirals that closely resemble the way a moth tracks a scent.
The algorithm uses information received from the scent itself as well as information received when the scent is not detected, striking a balance between heading directly toward the point where it guesses the scent is coming from and wandering around collecting information but not making any progress toward the source.
Vergassola says the algorithm could be implemented in an actual robot or be used for other applications that involve searching without much information, such as detecting the best paths for information to be sent through a network.
This research “provides a new framework for understanding a large and significant class of problems encountered in real world situations,” says the University of Pennsylvania’s Alan Gelperin. He adds that by adding instruments that could gather information about the airflow around a robot, the algorithm could even be improved.
Abstracted by the National Law Enforcement and Corrections Technology Center (NLECTC) from New Scientist (01/24/07); Inman, Mason.