An electronic nose (e-nose) uses a sensor array that reacts to air pollutants by leaving behind a ‘fingerprint’ that can be attributed to an odour. It is essentially composed of a sample delivery system, detection system and computing system. Common sensors include metal oxide semiconductor, conducting polymer sensors, acoustic wave sensors, field-effect gas sensors, pellistors, and fibre-optic sensors. The most commonly used sensors are conducting polymers, bulk acoustic devices and metal oxide semiconductor.
All electronic noses aim to provide odour concentration similar to that which is measured through olfactometry by analytical analysis of the chemical composition of the sample. However, it has been proven that correlation between analytical results and actual odor perception is not direct due to potential interactions between several odorous components.
The objective and main sought after benefit of e-noses is continuous feedback of “odour levels”. The main drawback of e-noses are the calibration phase. The non-linear synergy between different chemicals and the continuous sensor drift (loss of sensitivity due to age, temperature, humidity, etc.) requires a more advance calibration method than what has been historically implemented in electronic noses. Other technical issues including mold and contamination of the sample lines have been widely reported for most electronic noses available on the market.
The majority of development in electronic noses technology has been focused on development of more sensitive sensors. This approach ignores the fundamental challenge with machine olfaction which is the simulation of odour perception not detection. Human odour perception is the interpretation of the signals which are received from the olfactory bulbs and therefore relies heavily on the human brain. Any electronic nose that would mimic the human olfaction must also simulate the perception of teh odour which would require simulation of the entire human olfactory system.