The detection scheme presented in this Thesis utilizes threshold value to remove the void information in the fusion judgment layer of sensing system and process the date collected in the sensing layer with related methods and then rank based on time with fusion algorithm. Then the fusion method model of neural network is used and the algorithm is utilized for improvement. The experiment result shows that the scheme reduces the convergence time to a great extent, makes it effective application in the sensor with relatively few electric quantity, capable of improving the predication accuracy to a great extent, improving data monitoring efficiency and effectively reducing power consumption.
展开▼
