Influence of MOS Gas-Sensor Production Tolerances on Pattern Recognition Techniques in Electronic Noses

Krutzler, Christian; Unger, Astrid; Marhold, H.; Fricke, Thomas; Conrad, Thorsten; Schütze, Andreas

doi:10.1109/tim.2011.2161015

Cited by 29 publications

(9 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, the hierarchical evaluation strategy for underground fire detection presented in Sect. 5.3.2 was evaluated by Dekra/EXAM in the framework of the European project SAFETECH [109], and the potential for this approach and its certification was confirmed.…”

Section: Discussionmentioning

confidence: 97%

Sensor Arrays, Virtual Multisensors, Data Fusion, and Gas Sensor Data Evaluation

Reimann

Schütze

2013

Springer Series on Chemical Sensors and Biosensors

View full text Add to dashboard Cite

Selectivity will remain one of the main challenges for gas sensors. The majority of single gas sensors have a broad response spectrum and additional effort is required to provide specific information from the recorded raw data. Sensor arrays comprising a number of different sensors, either of the same basic type or combining different sensor principles, are often employed to improve the selectivity. Another approach is the dynamic operation of a single sensor element leading to a virtual multisensor. Depending on the application, both classification, i.e., recognition of a certain gas or gas mixture, and quantification, i.e., determining the concentration of a target gas, are required. For evaluation of the higher-dimensional data vectors obtained from the sensor system, different methods are commonly used, e.g., unsupervised and supervised multivariate statistics or artificial neural networks. For comprehensive classification, a multistep process is employed, typically comprising data preprocessing, feature extraction, dimensionality reduction, and finally classification. This chapter addresses typical methods used for gas sensor data evaluation and also briefly addresses challenges relating to sensor calibration and drift compensation. Four application examples are shown in detail demonstrating the use of different sensor configurations and data processing approaches: robust ozone quantification, sweat odor assessment trained by correlation with a human sensory panel, fire detection in buildings, and finally underground fire detection in coal mines requiring very high selectivity.

show abstract

Section: Discussionmentioning

confidence: 97%

Sensor Arrays, Virtual Multisensors, Data Fusion, and Gas Sensor Data Evaluation

Reimann

Schütze

2013

Springer Series on Chemical Sensors and Biosensors

View full text Add to dashboard Cite

show abstract

“…Good results were obtained in the qualitative and quantitative tests conducted using this instrument for three industrial gases: Acetone, chloroform, and methanol. Krutzler et al [34] showed that the determination of gas concentration is improved through the implementation of sensors with smaller resistance variations when the reference library from one sensor is also used for other sensor elements. In 2012, Brudzewski et al [35] used a nonlinear classifier in the form of a Gaussian kernel SVM to classify EN data.…”

Section: Related Workmentioning

confidence: 99%

Data Classification Methodology for Electronic Noses Using Uniform Manifold Approximation and Projection and Extreme Learning Machine

et al. 2021

View full text Add to dashboard Cite

The classification and use of robust methodologies in sensor array applications of electronic noses (ENs) remain an open problem. Among the several steps used in the developed methodologies, data preprocessing improves the classification accuracy of this type of sensor. Data preprocessing methods, such as data transformation and data reduction, enable the treatment of data with anomalies, such as outliers and features, that do not provide quality information; in addition, they reduce the dimensionality of the data, thereby facilitating the tasks of a machine learning classifier. To help solve this problem, in this study, a machine learning methodology is introduced to improve signal processing and develop methodologies for classification when an EN is used. The proposed methodology involves a normalization stage to scale the data from the sensors, using both the well-known min−max approach and the more recent mean-centered unitary group scaling (MCUGS). Next, a manifold learning algorithm for data reduction is applied using uniform manifold approximation and projection (UMAP). The dimensionality of the data at the input of the classification machine is reduced, and an extreme learning machine (ELM) is used as a machine learning classifier algorithm. To validate the EN classification methodology, three datasets of ENs were used. The first dataset was composed of 3600 measurements of 6 volatile organic compounds performed by employing 16 metal-oxide gas sensors. The second dataset was composed of 235 measurements of 3 different qualities of wine, namely, high, average, and low, as evaluated by using an EN sensor array composed of 6 different sensors. The third dataset was composed of 309 measurements of 3 different gases obtained by using an EN sensor array of 2 sensors. A 5-fold cross-validation approach was used to evaluate the proposed methodology. A test set consisting of 25% of the data was used to validate the methodology with unseen data. The results showed a fully correct average classification accuracy of 1 when the MCUGS, UMAP, and ELM methods were used. Finally, the effect of changing the number of target dimensions on the reduction of the number of data was determined based on the highest average classification accuracy.

show abstract

“…Technologies for gas sensors are rapidly developing due to the increasing demand for devices with characteristics compliant to a wide set of applications involving the gas monitoring challenge [1][2][3][4]. In particular, carbon dioxide (CO 2 ) measurement is a real need for several application contexts including environmental and agriculture monitoring.…”

Section: Introductionmentioning

confidence: 99%

Direct Printing of a Multi-Layer Sensor on Pet Substrate for CO2 Detection

et al. 2019

View full text Add to dashboard Cite

The use of inexpensive gas sensors is a real need for many applications requiring the use of disposable sensors. This work deals with the realization and characterization of a low cost CO2 sensor realized by rapid prototyping techniques. In particular, the sensor consists of a set of InterDigiTed electrodes, over which a double sensing layer made of PEDOT/PSS (CLEVIOS™ PHCV4, by H.C.Starck) and a solution of pristine graphene powder has been deposited. A silver nano-particle solution is used for inkjet printing the electrodes onto the PET (PolyEthylene Terephthalate) substrate, through a commercial inkjet printer. The sensing strategy is based on the variation of the electrical conductance of graphene due to gas molecules adsorption. The device responsivity observed in two different operating conditions (50 °C and 60 °C), is 4.0 µΩ/Ω/ppm and 4.7 µΩ/Ω/ppm. The corresponding values of the resolution are 400 ppm and 420 ppm. Main advantages of the developed sensor consist in the cost-effective fabrication techniques and the device flexibility, which are strategic for applications requiring disposable and shapeable devices to be installed into irregular surfaces.

show abstract

Influence of MOS Gas-Sensor Production Tolerances on Pattern Recognition Techniques in Electronic Noses

Cited by 29 publications

References 26 publications

Sensor Arrays, Virtual Multisensors, Data Fusion, and Gas Sensor Data Evaluation

Sensor Arrays, Virtual Multisensors, Data Fusion, and Gas Sensor Data Evaluation

Data Classification Methodology for Electronic Noses Using Uniform Manifold Approximation and Projection and Extreme Learning Machine

Direct Printing of a Multi-Layer Sensor on Pet Substrate for CO2 Detection

Contact Info

Product

Resources

About