A study on local sensor fusion of wireless sensor networks based on the neural network

Taken as the whole networkpsilas information fusion, local sensor fusion, integrating signals from different sources and processing locally, is the first step work. Due to the output of sensor nodes is vulnerable to the impact of around environmental factors, such as temperature, humidity, noise, etc., and in order to solve nonlinear problems between input and output, a sensor output compensation model based on the Neural Network is proposed. As the same time, the theory of the Neural Network is outlined, mainly an introduction is made to typical fusion algorithms, along with analyses and comparisons, in three Feed-Forward neural networks, BP, RBF and CMAC, respectively.