Olfactory Recognition and Synthesis: Does the Nose Know?
A digital dreamer’s goal is to duplicate the five human senses: touch, sight, hearing, smell and taste, ergo creating a robot that can do, as the song says, “anything you can do, I can do better.” (Unfortunately, sex hasn’t yet figured into the dream – yet.) Advances have been made in taste (there are only four), touch, hearing (especially voice) and sight. Wall-E and Artoo aren’t far off the reality mark.
The toughest sense to conquer appears to be that of smell. Most of you are too young to remember “Smell-O-Vision,” a 1960s attempt by Mike Todd’s son (Mike Todd was a famous movie producer, creator of Todd-A-O and husband to Liz Taylor) to add smells to motion picture viewing. Approximately 30 odors were synthesized and pumped into the theater at appropriate times. This early attempt was unsuccessful and was quickly abandoned by the movie industry.
The applications for synthetically duplicating the olfactory sense, however, are very important, and, to my way of thinking, justify a much larger research investment that is presently being made. Drug interdiction, security, disease detection, contaminant detection, gas leaks, crime prevention are but a few of the very important applications of the sense of smell. The best we have right now for these applications are dogs and vultures (for gas leaks)!
There are bazillions of molecules that can contribute to odors and more bazillions of permutations and combinations of those molecules. It would be nice if we could make up a table of these molecules and simply look them up, but that approach simply isn’t in the cards. The numbers are just too big.
A researcher in this field is Dr. Paul Rhodes. This Stanford Visiting Scholar formed a company called Evolved Machines that is attempting to use neural network technology to identify odors. Unfortunately, I don’t think he has gotten very far – his last press release was issued two years ago!
Given the enormous market potential of applications based on smell, it is a wonder that would-be entrepreneurs are not all over this one. The IEEE publishes a Sensors Journal that has scholarly articles published on the subject from time-to-time, but the apparent lack of serious research is astonishing to this observer. Do a Google search and be amazed at the paucity of informative hits on the subject.
So this subject remains an open question. Will it have a future? Maybe, even probably, but it looks like it will be a very long way off.