Intelligent Fiber Optic Statistical Mode Sensors Using Novel Features and Artificial Neural Networks

EFENDİOĞLU H. S., TOKER O., YILDIRIM T., FİDANBOYLU K.

SPIE Smart Structures (NDE 2013), 10 - 14 Mart 2013, cilt.8693, ss.869301-869306 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 8693
  • Doi Numarası: 10.1117/12.2009836
  • Sayfa Sayıları: ss.869301-869306
  • Bursa Uludağ Üniversitesi Adresli: Hayır

Özet

In this paper, intelligent statistical mode sensors are proposed and analyzed. Several statistical features are used in design of intelligent sensor systems. Force measurement experiments are conducted and experimental data is analyzed using newly proposed statistical features. After that, Artificial Neural Networks (ANNs) with sensor data fusion, which is an intelligent sensor architecture, was proposed to estimate the force values. Multilayer perceptron (MLP) with different algorithms are used in the ANN model, and all of them can predict the force values with acceptable error levels. Using sensor fusion with ANNs, statistical mode sensors can be calibrated and fault tolerance of the sensor can be decreased, hence more reliable intelligent sensors can be designed.