Abstract:

In the LTE-R handover of high-speed railway, the A3 event-based handoff algorithm is prone to the problem of Ping-Pong Effect（PPE） and Wireless Link connection Failure（WLF）, the Particle Swarm Optimization（PSO） gray wolf algorithm, the improved RBF neural network（IGWO-RBF） handover optimization algorithm is proposed. The algorithm firstly collects [Hys] and [TTT] parameter sets with high success rate when a large number of trains run in different environments at different speeds（0~100 m/s）, and sends them into the improved RBF neural network. After training, they get [Hys] at different speeds. The fitting curve with [TTT] is then based on the Received Signal Received Quality（RSRQ） of the train, and the self-correcting term is added to perform secondary optimization adjustments on [Hys] and [TTT]. Finally, the simulation experiments on matlab show that the proposed algorithm reduces the call drop rate and ping-pong switching rate, and improves the switching success rate and robustness of the train in high-speed operation environment.

Key words: LTE-R technology, high-speed environment, handover, A3 event, Improved Gray Wolf Optimizer RBF neural network（IGWO-RBF）, switching success rate

摘要：

针对高速铁路LTE-R越区切换中，A3事件下的越区切换算法容易出现乒乓效应（PPE）和无线链路连接失败（WLF）的问题，提出了粒子群优化（PSO）灰狼算法改进的RBF神经网络（IGWO-RBF）的越区切换优化算法。该算法采集大量列车以不同速度（0~100 m/s）运行在特定环境中时切换成功率高的切换迟滞门限[(Hys)]和触发延迟时间[(TTT)]参数集，送入改进的RBF神经网络，训练完成后得到不同速度下的[Hys]和[TTT]的拟合曲线。根据列车接收到的参考信号接收质量（RSRQ），加入自矫正项对[Hys]和[TTT]进行二次优化调整。在matlab上进行仿真实验，结果表明提出的算法减小了掉话率和乒乓切换率，提高了列车在高速环境下的切换成功率及鲁棒性。

关键词: LTE技术, 高速环境, 越区切换, A3事件, 改进灰狼优化的RBF神经网络（IGWO-RBF）, 切换成功率