Handover Enhancement in High-Speed Railway 5G Networks: A LSTM-based Prediction Method
Yuping Lu, Cuntao Zhang, Deen Chen, Wanle Zhang, Ke Xiong
Abstract
This paper focuses on the handover performance enhancement of high-speed railway (HSR) 5G millimeter wave(mmWave) networks. As starting the handover at different time may yield very different performances, we propose a Long Short Term Memory (LSTM)-based prediction method to find a proper handover point in advance to alleviate the high handover delay and link interruption-prone problems of traditional A3 handover method. By learning the historical trend of Reference Signal Receiving Power (RSRP) and predicting the future changes of RSRP based on the LSTM encoder-decoder network, the proposed LSTM-based prediction method is able to enhance the handover performance of HSR 5G networks. Simulation results show that compared with traditional A3 method, our proposed LSTM-based prediction method reduces the probability of link interruption, improve the stability of the quality of service (QoS) during the handover process.