International Journal of Automation and Computing 17(2), April 2020, 179-209 DOI: 10.1007/s11633-019-1212-9 Electronic Nose and Its Applications: A Survey Diclehan Karakaya Oguzhan Ulucan Mehmet Turkan Department of Electrical and Electronics Engineering, Izmir University of Economics, Izmir 35330, Turkey Abstract: In the last two decades, improvements in materials, sensors and machine learning technologies have led to a rapid extension of electronic nose (EN) related research topics with diverse applications. The food and beverage industry, agriculture and forestry, medi- cine and health-care, indoor and outdoor monitoring, military and civilian security systems are the leading fields which take great ad- vantage from the rapidity, stability, portability and compactness of ENs. Although the EN technology provides numerous benefits, fur- ther enhancements in both hardware and software components are necessary for utilizing ENs in practice. This paper provides an ex- tensive survey of the EN technology and its wide range of application fields, through a comprehensive analysis of algorithms proposed in the literature, while exploiting related domains with possible future suggestions for this research topic. Keywords: Artificial intelligence, machine learning, pattern recognition, electronic nose (EN), sensors technology. 1 Introduction aromas with a chemical electronic sensor array was primarily mentioned in [12] and then in [13] in the early All kinds of innovation are possible with inspiration. 1980s. However, the EN concept could not be actualized As image processing is inspired by the sense of sight, the at that time due to limitations in the sensors technology. electronic nose (abbreviation EN, enose, e-nose) – also In the late 1990s then, the term “electronic nose” was known as an odor sensor, aroma sensor, mechanical nose, mentioned in [14]. According to its initial definition, an flavor sensor, multi-sensor array, artificial nose, odor- EN is composed of a multisensor array responsible for de- sensing system, electronic olfactometry[1] – technology is tect- ing more than one chemical component. Sub- inspired by the sense of smell. The olfactory system eas- sequently, both technological improvements in sensors ily enables living beings to be aware of their environment, and the realization of the potential that the EN holds led of possible dangers, and to identify and classify food[2]. In to a considerable extension of its applications. Recently, technology though, automatic identification and classific- due to the provision of reliable solutions, rapidity, low ation of odor is a very challenging issue because the cost and compactness, the EN concept has become popu- scents in chemical mixtures intercommunicate naturally[3]. lar in agriculture[15], the food and water industry[16], medi- This natural interaction has three types: synergism, com- cine[17], security systems[15] and many other areas[18]. pensation and masking. Synergism is defined as the inter- Fig. 1 demonstrates the similarities between the biolo- action when two or more distinct substances produce a gical olfactory system and the EN technology. The elec- mutual scent which is stronger than those of individual tronic sensor array of the EN corresponds to the olfact- components. Compensation is the case when one compon- ory nose receptors, which detects the traces of chemicals ent counteracts another constituent. Masking is the com- in the air. When these molecules are sensed and captured, bination of one pleasant odor with an unpleasant one. the input signal is sent to the olfactory bulb, where the Even though there exist reported achievements of some odor information is processed. After characterizing the earlier techniques[4–8], such chemical mixtures in general aroma, the smell is recognized in the brain by the olfact- conditions were not able to be analyzed and split up into ory cortex as a last stage in the biological system. Like- its components with high accuracy until the development wise, in the EN case, a preprocessor applies feature ex- of the EN technology. Together with the development of EN devices, several studies were completed to assess odor Odor intensities, to understand mixtures of odor interactions particles and the sensor responses to these interactions, e.g., [9– 11]. Nose Brain Although the first research on detecting distinctive smells began in [4] during the 1920s, the idea to detect Sensor Data Review array analysis Manuscript received August 5, 2019; accepted November 15, 2019; published online December 28, 2019 Fig. 1 Analogy between the biological olfactory system and the Recommended by Associate Editor Jyh-Horng Chou EN technology © The Author(s) 2019 180 International Journal of Automation and Computing 17(2), April 2020 traction to the captured odor signal. Afterwards, data tained signals should be preprocessed to understand those analysis, pattern recognition and machine learning re- physical changes properly and then processed to digital- lated algorithms are generally used for identifying and ize them in order to form a dataset. Hence, the sensed classifying the input scent using the extracted digital sig- signals are appropriately manipulated, e.g., amplified, natures. filtered or converted, in order to be easily used in further After drawing the parallels between the biological ol- stages[22]. The processed signals are later analyzed in factory system and the EN technology, an EN obviously terms of their specific properties in the data gathering consists of both hardware and software components. stage. Subsequently, sufficient data is acquired from these While the software part can be thought as the “brain”, signals and the obtained data is preprocessed according to the hardware part can be seen as the “olfactory recept- the requirements of the employed pattern recognition al- ors” of the EN system. The software part mainly con- gorithm. Lastly, the odor is classified with the pattern re- tains a data processing unit which identifies and classi- cognition stage. fies each individual scent detected using digital signa- tures of the sensed chemicals. The hardware part is basic- Odor Sensor Signal Signal ally a sensor array. Since the main objective of the EN is release array conditioning processing detecting and classifying multiple aromas, the sensing ar- ray should encompass different types of individual sensors, where each sensor is responsible for detecting a Odor Pattern Data Data classification recognition preprocessing gathering different chemical. The selection of suitable sensors to a given specific task is a key point in this technology. It can Fig. 2 A description of EN including both hardware and be concluded here that choosing the suitable hardware software components and efficient software components is very important in designing and implementing a successful EN system for a 2.1 Sensors and chemicals particular problem. Hence, the main aim of this survey is to provide a complete overview of the EN technology The olfactory system can neither detect nor identify while pointing out the task-dependent importance of its scents without the obtaining of chemicals released by the hardware and software components and their character- objects[23]. These chemicals can be found in simple or istic properties. Note that there are several existing re- complex structures. Nevertheless, each individual chemic- views in the literature which focus on a specific sub-topic al has its own unique quality and characteristics. Thus, of the EN concept, e.g., the range of sensors as hardware digital signatures of chemicals, which are to be the input [19] components used in the EN systems , ENs for the food data for both the olfactory system and the EN instru- [20] industry , neural networks as software components for ment, are exclusive. The automatic detection process can- ENs[21]. On the other hand, this paper rather provides a not be achieved without a collected library (dataset) of comprehensive review of broad EN application fields, a the digital signatures of specific aromas. wide range of software related algorithms and commonly The sensor array is responsible for detecting targeted utilized sensor types and their properties as practical chemicals in a medium. Each targeted aroma is detected hardware components. with a specific sensor, in other words each individual The remaining part of this survey is structured as fol- sensor is responsible for sensing a specific type of aroma. lows. The EN device and its main components are intro- Chemical sensors are used for detecting chemicals in the duced in Section 2. Then, a wide range of practical EN medium. These sensors basically convert chemical inform- applications and related topics are described in Section 3. ation into analytical signals[24]. Afterwards, current challenges for this technology are dis- Since the main objective of an EN is to sense more cussed in Section 4, followed by perspectives and possible than one chemical, this aim can be achieved with higher future research directions in Section 5. This paper is fi- accuracy only by combining several distinct sensors in the nally concluded with a brief conclusion in Section 6. array. While keeping in mind that the sensors in the ar- ray have to be chosen carefully by taking the chemicals of 2 Components of an EN interest into account[25], having a proper sensor array for An EN consists of both hardware and software com- specific tasks depends on several conditions as follows[26]: ponents as briefly depicted in an