Data Fusion Algorithm based on Functional Link Artificial Neural Networks

Abstract

Data fusion is the process of combining data from several sources into a single unified description of a situation. The sensor output value is estimated by some data fusion algorithms. In this paper, functional link artificial neural networks (FLANN) and data fusion technique are combined for removing the ambient temperature disturbance to enhance accuracy and reliability of lumber moisture content sensors (LMCS). Three different functional expansions, Chebyshev, Legendre and power series are studied. Simulation results show that the performance of Chebyshev polynomials is superior to the other two FLANN model. Compared with MLP, C-FLANN exhibits a much simpler structure, less training computation and faster convergence. So it is easier to implement by hardware and improve the performance-price ratio of system. The experimental results show that FLANN data fusion method can eliminate effectively the measurement errors and get reliable, real-time, accuracy estimated output of sensor