Wireless Channel Modelling And Forecasting Of Semicircle Arch Tunnel Based On LS-SVM

As technology evolving,mine accident rate has been reduced to a great extent.However,the safety production problems are still serious and mine accident rescue still lacks effectiveness.A reliable and sensitive underground wireless communication system is needed badly.Due to the complex underground environment,propagation of the electromagnetic wave is affected by refraction and scattering from tunnel structure,coal seam and locomotives,and electromagnetic interference from high-power devices.The fading of electromagnetic wave is complicated.Thus,the prediction of mine wireless channel lacks accuracy,limiting the development of underground wireless communication technology.On the basis of studying the theory of electromagnetic wave propagation,least squares support vector machine(LS-SVM)is used to establish the prediction model of underground tunnel wireless channel,which provides guidance to designing underground wireless communication system.Firstly,relevant theory knowledge about support vector machine is introduced.After analyzing mine wireless channel,the propagation mode of electromagnetic wave is studied.The attenuation of electromagnetic wave is affected by many factors including the frequency of the antenna,the parameters of the coal seam,the roughness and tilt of the tunnel wall and cross section shape of the tunnel.Modeling and prediction of path fading loss and multi-path fading loss in semi-circular arch tunnels has been studied.The factors that affect the propagation of electromagnetic wave are used as input variables to establish the LS-SVM prediction model and the path loss is the output.In the model training,the genetic algorithm is used to optimize penalty factor and kernel function parameter in the LS-SVM algorithm.Compared with wavelet neural network and LS-SVM of grid search,the effectiveness of the algorithm is verified.For mult-path fading,the Nakagami fading channel is studied.Nakagami fading channel is simulated to achieve the data sequence of channel fading coefficient.Principal components analysis is used to preprocessing data and phase space reconstruction theory is used to reconstruct the fading coefficients.Based on the restructed data,LS-SVM model is established.The simulation results show that the prediction model based on LS-SVM has a good performance of predicting the change of channel.Compared with traditional prediction algorithms,MAPE,RMSE and correlation coefficient between predicted data and original data with improved SVM can be improvd and it has higher precision and better effect.It is proved that support vector machine algorithm is a feasible method to model the electromagnetic wave propagation.