Sniffer Robots: Artificial Noses

1
Sniffer Robots Artificial Noses

• Brandon Yarborough

2
Components of an Artificial Nose

• Chemical sensing system (e.g., sensory array or
  spectrometer)
• Each chemical vapor presented to the sensor array
  produces a pattern characteristic (signature) of
  the vapor.
• A database of signatures is built up by
  presenting many different chemicals to the sensor
  array.
• The database of labeled signatures is used to
  train the pattern recognition system.

3
Components of an Artificial Nose (cont.)

• Pattern recognition system (e.g., artificial
  neural network)
• Used to analyze complex data and to recognize
  patterns
• Artificial neural networks (ANNs) are collections
  of mathematical models that attempt to emulate
  some of the properties of biological nervous
  systems.

4
General Structure of an Artificial Nose
5
Possible Applications for Sniffer Robots

• Detecting drugs and other materials in airports
• Locating victims of avalanches and earthquakes
• Mine/bomb detection
• Detection of chemical leaks
• Tracking escaped prisoners

6
RAT The Reactive Autonomous Testbed

• Created by Andrew Russell at Monash University in
  Melbourne, Australia
• The smellbot, as it is known, can successfully
  detect and follow odors through a maze.
• Can detect odors with a concentration of as
  little as one part per million (though this is
  not as efficient as a dog, or even a human)

7
Pictures of RAT
8
Videos of RAT in action

• http//www.ecse.monash.edu.au/staff/rar/Trail202
  31.mov
• http//www.ecse.monash.edu.au/staff/rar/Plume202
  31.mov

9
Sources

• http//www.newscientisttech.com/article/dn11023-sn
  ifferbot-algorithm-helps-robots-seek-scents.html
• http//www.ecse.monash.edu.au/staff/rar/
• http//www.theage.com.au/articles/2003/09/10/10629
  02115129.html?fromstoryrhs