Can Online Learning Increase the Reliability of Extreme Mobility Management?

Abstract

Seamless Internet access under extreme user mobility is highly demanded on high-speed trains and vehicles. However, existing mobile networks (e.g., 4G LTE and 5G NR) cannot reliably satisfy this demand, with a 5.5%-12.6% handover failure ratio at 200–350 km/h. A root cause is that, the 4G/5G handovers have to balance the exploration of more measurements for satisfactory handover and the exploitation for timely handover before the fast-moving user leaves the coverage.We design BaTT, an online learning solution for reliable handovers in extreme mobility. BaTT decomposes the explorationexploitation tradeoff into two multi-armed bandit problems. It uses ϵ-binary-search to optimize the threshold of a serving cell’s signal strength to initiate the handover with $\\mathcal{O}(\\log J\\log T)$ regrets. It further adopts opportunistic Thompson sampling to optimize the sequence of target cells measured for reliable handovers. BaTT can be implemented using the recent Open Radio Access Network (O-RAN) framework in operational 4G LTE and 5G NR. Our evaluations over a dataset from operational LTE networks on the Chinese high-speed rails show a 29.1% handover failure reduction at the speed of 200-350 km/h.