Handoff Algorithm Of High Speed Railway Wireless Communication Based On Neural Network

The emergence of China’s high-speed railway not only improves social and economic benefits,but also facilitates passengers’ travel,which brings certain opportunities to cities along the railway.In recent years,high-speed rail technology has been continuously optimized and improved.High-speed railway wireless communication is an important technology in the construction of high-speed railway,which ensures the reliable transmission of train dispatching information,realizes the communication between high-speed running trains and the ground,and meets the needs of passengers on the Internet.When the train runs from one base station to another,handoff will inevitably occur in order to ensure the communication between the train and the ground without interruption.When switching,the possibility of switchover interruption should be reduced,the ability of system resource scheduling should be strengthened,and the success rate of switchover should be increased.However,in the handover of high-speed railway wireless communication,the high-speed train running at high speed will lead to insufficient handover time,the generalization ability of the algorithm under complex and changeable conditions is poor.Therefore,it is necessary to study the handover of railway wireless communication.To solve the above problems,this dissertation proposes a dual-link handover algorithm based on CNN-WaveNet decision parameter multi-step prediction model and a LTE-R adaptive handover algorithm based on beamforming.The main work and research results are as follows:(1)Aiming at the problem that the train speed increases,the train stay time in the overlapping area decreases,and the probability of successful handover is low.A dual-link handover algorithm based on CNN-WaveNet decision parameter multi-step prediction model is proposed.CNN is used to extract the time sequence features of decision parameters,and then WaveNet is used to learn the mapping relationship between feature information and decision parameters,and combined with prediction,multi-step prediction is realized.According to the predicted values,resources can be switched in advance.At the same time,dual-antenna communication is used to realize dual-link communication.When the current antenna fails to switch,another antenna can be switched once.The purpose of this algorithm is to improve the prediction accuracy of the model,and at the same time strengthen the ability of system resource scheduling,and improve the handover trigger probability and handover success probability.Compared with the single-link handover algorithm without predictive model,the proposed dual-link handover algorithm based on decision parameter prediction model can effectively improve the handover trigger probability and handover success probability.(2)Aiming at the problem of poor generalization ability of complex and changeable handover algorithm for high-speed railway and low handover trigger probability.An adaptive LTE-R handover algorithm based on beamforming is proposed.The algorithm uses neural network to predict the received power.If the predicted received power can not guarantee good communication,the beamforming technique is activated.Considering that the intensity of each base station will change when the train is running,a dynamic function is established between the train position and the beam gain value to dynamically dynamic the beam gain value according to the train position information.In addition,the switching threshold is adaptively optimized based on fuzzy logic,considering the received power,signal to interference ratio,train position and train running speed.The generalization ability of the algorithm can be improved by dynamic beam gain value and switching threshold of train reported information.Simulation results show that the LTE-R adaptive handover algorithm based on beamforming proposed in this dissertation is compared with the handover algorithm using fixed handover threshold.The proposed algorithm improves the handover success probability and handover trigger probability.