Low cost artificial noses with communication and collaboration capabilities
Author
Urízar Salinas, MiguelAdvisor
Varona Martínez, PabloEntity
UAM. Departamento de Ingeniería InformáticaDate
2014Subjects
Detectores; Olfato; Olores; InformáticaNote
Máster Universitario en Investigación e Innovación en Inteligencia Computacional y Sistemas InteractivosEsta obra está bajo una licencia de Creative Commons Reconocimiento-NoComercial-SinObraDerivada 4.0 Internacional.
Abstract
An artificial nose sensor network has been successfully installed and used for monitoring a biodiesel plant for two years. The network is composed of different low-cost artificial noses distributed in several locations of the factory.
Changing or adding noses during this time was possible, demonstrating that the system is modular and flexible. Communications stabilized by serial cables have been operating continuously without failure, as it has also made the data acquisition program except for some outages, being these external to the system.
This work describes the main features found in the signals recorded by the artificial noses during loading task, leak and routine events, including inactivity periods. A set of automatic classification experiments based in support vector machine (SVM) classifiers (a generic one (SVM), a cross validation SVM (CSVM), an autoregressive kernel SVM (ARSVM) and a dynamic systems kernel SVM (DSVM)) have been developed for five different event types at two sampling rates. Different variables depending on the classifier have been sweep to find their best values: for SVM gamma, cost and tolerance and for CSVM ARSVM and DSVM only tolerance. Artificial noses have noise, drift and also history-dependent sensitivity, so the defined environment is complex. Nevertheless, for almost all events, the classifiers reach a 75% of detection or higher. The results reported in this work show that these noses can detect leak and routine events in the plant. Furthermore, continuous monitoring seem to indicate that low-cost artificial noses can be used to characterize the activity of the plant in a non-invasive manner, suggesting new uses for machine olfaction.
Files in this item
Google Scholar:Urízar Salinas, Miguel
This item appears in the following Collection(s)
Except where otherwise noted, this item's license is described as https://creativecommons.org/licenses/by-nc-nd/4.0/
Related items
Showing items related by title, author, creator and subject.
-
Artificial cognitive architecture with self-learning and self-optimization capabilities. Case studies in micromachining processes
Beruvides López, Gerardo
2017-09-22 -
Communication strategies in Spanish academic libraries
Albertos, Ana; Culebras, Javier; Ordi, Isabel del; Sánchez Salinas, Rosario
2014-03 -
AI4FoodDB: a database for personalized e-health nutrition and lifestyle through wearable devices and artificial intelligence
Romero Tapiador, Sergio; Lacruz-Pleguezuelos, Blanca; Tolosana Moranchel, Ruben; Freixer, Gala; Daza Garcia, Roberto; Fernández-Díaz, Cristina M.; Aguilar-Aguilar, Elena; Fernández-Cabezas, Jorge; Cruz-Gil, Silvia; Molina, Susana; Crespo, Maria Carmen; Laguna, Teresa; Marcos-Zambrano, Laura Judith; Vera Rodriguez, Rubén; Fiérrez Aguilar, Julián; Ramírez De Molina, Ana; Ortega Garcia, Javier; Espinosa-Salinas, Isabel; Morales Moreno, Aythami; Carrillo De Santa Pau, Enrique
2023-07-18