Intelligent Soft Sensors

• Format: eBook
• Language: English
• Published: MDPI - Multidisciplinary Digital Publishing Institute 2023
• Subjects: History of engineering and technology; Technology: general issues; affective computing; bioprocess monitoring; BIS index; computerized adaptive testing (CAT); D-S evidence theory; depth of hypnosis; early fire warning; EDA; electrical resistance; executive functions; extended Kalman filter; extreme learning machine; frequency analysis; general anesthesia; hybrid feature fusion; image feature extraction; improved mathematical model; improved particle swarm algorithm; intelligent building system; joule heating effect; keyframe extraction; kinetic model; least squares support vector machine; modelling; multi-source data fusion; n/a; neurodevelopmental disorders; non-linear models; nonlinear regression model; nonlinear systems; observability; outliers; physiological signals; Pichia pastoris; population-data-based model; prognostic and health management; propofol; Raman; residual model; robust observer; self-sensing actuation; sensor selection; shape memory coil; simulator; sintering quality prediction; soft sensor; soft sensors; soft-sensor based diagnosis; spectroscopy; state estimation; stress detection; support vector machine regression model; target-controlled infusion; total intravenous anesthesia; transfer learning; variable selection; variable stiffness actuation
• Online Access: Open Access: DOAB: description of the publication; Open Access: DOAB, download the publication

 This Special Issue deals with the field of intelligent soft sensors that enable the online estimation of nonmeasurable process variables. Soft sensors or virtual sensors are common names for software algorithms in which multiple measurements are processed together. Typically, soft sensors are based on control theory and are also referred to as state observers. There may be dozens or even hundreds of measurements from hard sensors (big data). The interaction of signals can be used to compute new quantities that cannot be measured directly online or are difficult and expensive to measure. Soft sensors are particularly useful in data fusion, combining measurements of different characteristics and dynamics. They can be used for fault diagnosis (self-analysis, self-calibration, and self-maintenance) as well as for control applications. Well-known software algorithms that can be seen as soft sensors include, for example, Kalman filters. More recent implementations of soft sensors use neural networks, fuzzy logic, models based on evolving clustering, partial least squares, etc. In the digitized factories of the future, intelligent sensors represent one of the core building blocks for automating and optimizing production, as they make production more efficient in every respect.