Prediction-Based Proactive Handover for Reliable Connectivity in mmWave and Terahertz Vehicular Networks

초록

Millimeter-wave (mmWave) and terahertz (THz) communications are key enablers for high-capacity, low-latency vehicular networks, but their vulnerability to blockage and rapid signal fluctuation poses critical challenges for mobility management. In particular, conventional reactive handover (HO) schemes often fail to respond in time, leading to radio link failures (RLF) or unstable connections. This paper proposes a prediction-based proactive HO framework that forecasts future Reference Signal Received Power (RSRP) values using a Transformer-based model and makes anticipatory HO decisions over a sliding prediction window. The proposed method includes a dual-branch decision logic that either optimizes link quality or prevents imminent degradation. Extensive simulations under both static and vehicle-induced blockage scenarios compare the proposed method against conventional reactive handover and existing learning-based approaches. Results show that the proposed approach achieves a balance across handover failure metrics, reducing Too-Late Handover (TLHO), Too-Early Handover (TEHO), and RLF events, while improving throughput and connection stability. These results highlight the potential of learning-based proactive HO strategies for enhancing reliability in mmWave and THz vehicular networks. © 1967-2012 IEEE.

키워드