TruffleBot Sniffs Out and Accurately Identifies Specific Chemical Odors

Electronic noses have been around for a while, and the general public is probably most familiar with them in breathalyzers used to test blood-alcohol levels. But, they’re used in all kinds of other industries as well. The vast majority of them work with chemical sensors that have some kind of electronically-detectable change in the presence of the specific chemical they were designed to look for. TruffleBot takes that a step further and physically “sniffs” for a variety of compounds.

TruffleBot was created by Jacob Rosenstein, who is an electrical engineer at Brown University. Rosenstein and his team were inspired by the way animals sniff out specific odors. They don’t rely just on air passively moving across the odor receptors in their noses, they actively inhale that air to identify the smells they’re looking for. TruffleBot works in the same way by using small air pumps to push air across its sensors.

In total, TruffleBot has 16 sensors placed on a Raspberry Pi HAT. Eight of those are traditional electronic nose chemical sensors, and the other eight are mechanical digital barometers that monitor air pressure and temperature. Each chemical sensor is paired with a mechanical sensor, and the pairs are arranged in four rows of two pairs each. In testing, TruffleBot was used to analyze nine odors, and could do so with 80% accuracy using the chemical sensors alone. With the air pumps going, that jumped to 90%, and then all the way to 95% with the mechanical sensors.