Breakthrough in early safety warning system for civil aviation airport based on BP neural network
Ke Pan · Applied Mechanics and Materials · 2008
This paper is aimed to report its authors' breakthrough renovation over the conventional early warning model of airport accidents by using the BP network system in its process.In our research,first of all,we have defined more clearly the characteristic features of BP network and pointed out the out-of-date operational routine still prevailing in the current Chinese airports,thus,laying a successful foundation to its application.In doing so,we have also explained why it is necessary to choose the two levels of the BP network for its practical application.Next,we have ascertained the index system of the early safety warning practice with the simulation data concerned being done by SPSS principal components analytic method.Furthermore,the training and test of this BP model has been checked by using the MATLAB software.Test results of our research prove that our model is feasible and efficient for its mission.For practical purpose,we have also presented the structure,the function,as well as the principles of our renovated early warning system in a detailed way,which is highly applicable on the basis of bi-level constitutional model.