A large scale virtual gas sensor array

Andrey Ziyatdinov, Eduard Fernández-Diaz, A. Chaudry, Santiago Marco, Krishna Persaud, Alexandre Perera

    Research output: Contribution to journalArticlepeer-review


    This paper depicts a virtual sensor array that allows the user to generate gas sensor synthetic data while controlling a wide variety of the characteristics of the sensor array response: arbitrary number of sensors, support for multi-component gas mixtures and full control of the noise in the system such as sensor drift or sensor aging. The artificial sensor array response is inspired on the response of 17 polymeric sensors for three analytes during 7 month. The main trends in the synthetic gas sensor array, such as sensitivity, diversity, drift and sensor noise, are user controlled. Sensor sensitivity is modeled by an optionally linear or nonlinear method (spline based). The toolbox on data generation is implemented in open source R language for statistical computing and can be freely accessed as an educational resource or benchmarking reference. The software package permits the design of scenarios with a very large number of sensors (over 10000 sensels), which are employed in the test and benchmarking of neuromorphic models in the Bio-ICT European project NEUROCHEM. © 2011 American Institute of Physics.
    Original languageEnglish
    Pages (from-to)151-152
    Number of pages1
    JournalAIP Conference Proceedings
    Publication statusPublished - 2011


    • Data analysis benchmarking
    • Drift
    • Gas sensor
    • Sensor array


    Dive into the research topics of 'A large scale virtual gas sensor array'. Together they form a unique fingerprint.

    Cite this