【摘要】 分析了传统高铁LTE-R网络建设的业务需求。针对列车行驶中频繁出现的A3切换事件，此事件使得专网小区出现数据上传、下载速率低，话音接通率、掉线率劣化等问题，提出了一种基于改进的人工蜂群算法和径向基函数神经网络的越区切换优化方法。通过分析高铁列车的实时运行速度和传输损耗，对列车运行中切换触发时延和切换迟滞门限这两个参数进行了预测和动态优化，以提升高铁列车在运行中用户终端的切换成功率和鲁棒性。实验中以长吉线高铁某区间为例进行测试，采集列车在不同运行速度下的实际参数进行训练。实验结果表明，基于改进的工蜂群算法和神经网络的切换方法具有更高的实时性和稳定性，能够大幅提升用户业务感知。更多还原

【Abstract】 This paper analyzes the business requirements of traditional high-speed rail LTE-R network construction.In view of the frequent occurrence of A3 handover events during train operation, which makes the private network community have problems such as low data upload and download rates, deterioration of voice connection rate and drop rate, etc.This paper proposes a handover optimization method based on Improved Artificial Bee Colony Algorithm and Radial Basis Function Neural Network.By analyzing the real-time running speed and transmission loss of high-speed trains, the two parameters of handover trigger delay and handover hysteresis threshold during train operation are predicted and dynamically optimized to improve the handover success rate and robustness of user terminals during high-speed train operation.In the experiment, a certain section of the Changji high-speed railway is used for testing, and the actual parameters of the train at different operating speeds are collected for training.The experimental results show that the handover method based on the Improved Artificial Bee Colony Algorithm and Neural Network has higher real-time performance and stability, and can greatly improve the user’s business perception.更多还原