Title: New improved FLANN approach for dynamic modelling of sensors

Authors: Song-ling Huang, Kuan-sheng Hao, Wei Zhao

Addresses: Department of Electrical Engineering, Tsinghua University, West Main Building 1-313, Beijing, 100084, China. ' Department of Electrical Engineering, Tsinghua University, West Main Building 1-313, Beijing, 100084, China. ' Department of Electrical Engineering, Tsinghua University, West Main Building 1-313, Beijing, 100084, China

Abstract: An improved functional link artificial neural networks (FLANN) is presented and applied to the dynamic modelling for sensors. Compared with the traditional FLANN method, the improved FLANN method differs greatly in the calculation of partial derivatives of the weighting parameters and the dynamic model|s output because the dependence of the past dynamic model|s output on the parameters is fully considered. Thus, a more accurate evaluation of the gradient and a faster iteration process can be obtained. The simulation and experimental results of the infrared temperature sensor|s dynamic modelling show that the improved FLANN method has higher convergence rate and more robustness.

Keywords: temperature sensors; infrared sensors; dynamic modelling; identification; sensor modelling; functional link ANNs; artificial neural networks; FLANNs; simulation.

DOI: 10.1504/IJCAT.2011.042224

International Journal of Computer Applications in Technology, 2011 Vol.41 No.1/2, pp.4 - 10

Published online: 01 Sep 2011 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article