A two dimensional intelligent calibration of an ion sensor

• Published in: 1996 IEEE Instrumentation and Measurement Technology Conference Proceedings Vol. 2; pp. 788 - 791 vol.2
• Main Authors: Attari, M., Heniche, M., Boudjema, F.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1996
• Subjects: Artificial intelligence; Artificial neural networks; Backpropagation algorithms; Calibration; Instruments; Intelligent sensors; Multi-layer neural network; Neural networks; Sensor phenomena and characterization; Sensor systems
• ISBN: 0780333128; 9780780333123
• DOI: 10.1109/IMTC.1996.507276

This paper focuses on the use of a multilayered neural network as an effective tool in the instrumentation field. The emphasis is on the linearization of nonlinear sensors' characteristics in the presence of disturbances. An artificial neural network reproduces the inverse sensor's characteristic so that the global system (sensor and ANN) operates like a linear one. Two algorithms widely known in artificial neural networks theory were jointly used to adapt the weights during training, namely backpropagation (BP) and random optimization method (RO). In order to have a good illustration, an ion selective electrode has been used because of its high sensitivity to interfering ions present in a solution of interest. Accuracy curves including the disturbing variable are shown to discuss the performance of this method.