Double Threshold Handover Algorithm for High-Speed Railway Signaling Based on LSTM

Abstract During the operation of high-speed trains, handovers between base stations are necessary to ensure continuous communication. However, traditional handover algorithms are no longer adequate for the needs of high-speed trains. Therefore, a dual-threshold handover algorithm based on LSTM prediction is proposed to improve communication quality. Firstly, two size thresholds are set. The Long Short-Term Memory (LSTM) neural network is used to predict the handover hysteresis threshold parameters. If the predicted value is less than the smaller threshold, it is replaced by the predicted value; if the predicted value is greater than the larger threshold, it is replaced by the predicted value. The in-depth analysis of the experimental results further elucidates the remarkable advantages of the proposed LSTM based dual-threshold handover algorithm for high-speed rail signal systems. Specifically, this algorithm not only significantly enhances the success rate of handovers but also effectively reduces the ping-pong handover rate, which is crucial for ensuring communication continuity and stability in high-speed mobile environments. This not only alleviates network load but also minimizes communication interruptions and delays caused by frequent handovers, thereby notably improving passengers' communication experience quality.