Abstract. "Industrie 4.0" or the Industrial Internet of Things (IIoT) are two terms for the current (r)evolution seen in industrial automation and control. Everything is getting smarter and data generated at all levels of the production process are used to improve product quality, flexibility, and productivity. This would not be possible without smart sensors, which generate the data and allow further functionality from self-monitoring and selfconfiguration to condition monitoring of complex processes. In analogy to Industry 4.0, the development of sensors has undergone distinctive stages culminating in today's smart sensors or "Sensor 4.0". This paper briefly reviews the development of sensor technology over the last 2 centuries, highlights some of the potential that can be achieved with smart sensors and data evaluation, and discusses success requirements for future developments. In addition to magnetic sensor technologies which allow self-test and self-calibration and can contribute to many applications due to their wide spectrum of measured quantities, the paper discusses condition monitoring as a primary paradigm for introducing smart sensors and data analysis in manufacturing processes based on two projects performed in our group.

Smart sensors with internal signal processing and machine learning capabilities are a current trend in sensor development. This paper suggests a set of complementary and automated algorithms for feature extraction and selection to be used with smart sensors. The suggested methods for feature extraction can be applied on smart sensors and are capable of extracting signal characteristics from signal shape, time domain, time-frequency domain, frequency domain and signal distribution. Feature selection subsequently is capable of selecting the most important features for linear and nonlinear fault classification. The paper also highlights the potential of smart sensors in combination with the suggested algorithms that provide both data and further functionality from self-monitoring to condition monitoring in industrial applications. The first example applications are condition monitoring of a complex hydraulic machine where smart signal processing allows classification and quantification of four different fault scenarios. Additionally redundancies in the systems were used for self-monitoring and allowed to detect simulated sensor faults before they become critical for fault classification. The second example application is remaining lifetime prediction of electromechanical cylinders that shows applicability to big data and transparency of the solution by providing detailed information about sensor significance.

The classification of cyclically recorded time series plays an important role in measurement technologies. Example use cases range from gas sensors combined with temperature cycled operation to condition monitoring using vibration analysis. Before machine learning can be applied to high dimensional cyclical time series data dimensionality reduction has to be performed to avoid the classifier suffering from overfitting and the “curse of dimensionality”. This paper introduces a set of four complementary feature extraction methods and three feature selection algorithms that can be applied in a fully automatized manner to reduce the number of dimensions. The feature extraction algorithms are capable of extracting characteristic features from cyclical time series catching information contained in local details and overall cycle shape as well as in frequency or time-frequency domain. The methods for feature selection are capable of selecting the most suitable features for linear and nonlinear classification. The methods were chosen to be applicable to a wide range of applications which is verified by testing the set of methods on four different use cases.

With air quality being one target in the sustainable development goals set by the United Nations, accurate monitoring also of indoor air quality is more important than ever. Chemiresistive gas sensors are an inexpensive and promising solution for the monitoring of volatile organic compounds, which are of high concern indoors. To fully exploit the potential of these sensors, advanced operating modes, calibration, and data evaluation methods are required. This contribution outlines a systematic approach based on dynamic operation (temperature-cycled operation), randomized calibration (Latin hypercube sampling), and the use of advances in deep neural networks originally developed for natural language processing and computer vision, applying this approach to volatile organic compound measurements for indoor air quality monitoring. This paper discusses the pros and cons of deep neural networks for volatile organic compound monitoring in a laboratory environment by comparing the quantification accuracy of state-of-the-art data evaluation methods with a 10-layer deep convolutional neural network (TCOCNN). The overall performance of both methods was compared for complex gas mixtures with several volatile organic compounds, as well as interfering gases and changing ambient humidity in a comprehensive lab evaluation. Furthermore, both were tested under realistic conditions in the field with additional release tests of volatile organic compounds. The results obtained during field testing were compared with analytical measurements, namely the gold standard gas chromatography mass spectrometry analysis based on Tenax sampling, as well as two mobile systems, a gas chromatograph with photo-ionization detection for volatile organic compound monitoring and a gas chromatograph with a reducing compound photometer for the monitoring of hydrogen. The results showed that the TCOCNN outperforms state-of-the-art data evaluation methods, for example for critical pollutants such as formaldehyde, achieving an uncertainty of around 11 ppb even in complex mixtures, and offers a more robust volatile organic compound quantification in a laboratory environment, as well as in real ambient air for most targets.

No abstract