Research On Handover Algorithm Of LTE-R Wireless Communication In High-speed Railway

With the rapid development of high-speed railways,the requirements for railway mobile communication systems are becoming higher and higher.Advanced information and communication technology is the key to ensuring that high-speed trains are safe,efficient,comfortable and environmentally friendly,while also being able to provide passengers with convenient broadband communication services.GSM-R is a second-generation mobile technology,based on narrowband communication,which cannot meet the growing demand for upper layer applications.LTE-R is a railway broadband mobile communication system based on long-term evolution technology,which has a flat and IP-based network structure,can support a wide range of system bandwidths,and has a high communication rate to support high-speed movement,and can better meet the upper layer services of the high-speed railway mobile communication system.The reliability and efficiency of LTE-R depends on the effectiveness of its handover technology.When high-speed trains traverse the cell,the quality of service is affected by the interruption of vehicle-to-ground communication due to handover failure.Therefore,it is important to propose efficient handover algorithms for high-speed railway scenarios to reduce the handover failure rate and thus ensure the safety of high-speed railway operations and the evolution of GSM-R to LTE-R.In this paper,the following innovative work has been done from the perspective of improving the effectiveness of the handover technique:A fuzzy logic-based adaptive handover algorithm was proposed to address the problem of fixed hysteresis thresholds for handover,considering the influence of reference signal reception power,reference signal reception quality and speed on handover.The fuzzy logic inference mechanism was introduced into the handover algorithm,and the train speed,reference signal reception power and quality were used to optimize the hysteresis threshold adaptively.The results show that the proposed algorithm reduces the ping-pong handover rate near the midpoint of the handover overlap by 55.8% compared with the conventional A3 algorithm,effectively improving the effectiveness and reliability of handover.To address the problem that the traditional handover algorithm ignores the spatio-temporal correlation between measurement data,an LSTM recurrent neural network handover algorithm was proposed,which made use of the memory characteristics of LSTM neural networks and the spatio-temporal correlation of signals in the handover overlap area of high-speed railways to build a deep learning network for dynamic prediction of handover hysteresis parameters based on LSTM recurrent neural networks.The adaptive prediction of the handover hysteresis parameters is achieved,and the prediction error was kept between0.02.The adaptive prediction of the hysteresis parameters was achieved,and the handover performance was improved.To address the problem that the traditional A3 event-based judgement algorithm is difficult to adapt to the different handover requirements of control-plane handover and user-plane handover.An interval type two feature selection recurrent fuzzy neural network(T2RFS-FNN)based handover optimization algorithm for crossing zones was proposed.A parameter optimization model based on fuzzy neural network was established,and the convergence performance of the model was optimized by using an interval type II Gaussian affiliation function,while a feature selection layer was added to determine the output of the affiliation function to complete the optimization of the hysteresis threshold.The proposed algorithm achieves ping-pong handover rate below 0.645% for both user-plane handover and control-plane handover,and improved the success rate of handover.