In this paper, the information value model (IVM) of an access point (AP) and the related intelligent handover model (IHM) are firstly founded in the subway light-fidelity (LiFi) network. At first, based on statistical methods, the IVM is created by analyzing the behavioral characteristics of subway passengers. By the IVM, the probability of user access to each AP can be successfully predicted. Next, based on this probability, the IHM is created by deeply analyzing the alternating blockage problem of handholds. By the IHM, the system can self-adapt to the hysteresis state. This not only effectively avoids ping-pong handovers, but also reduces outage probability. Finally, by simulation, the effectiveness of the above IVM and IHM has been demonstrated. The simulation results indicate that the accuracy of predicting user target AP can be up to 92% by the IVM. Specially, compared with using the standard handover model, the throughput of the network not only increases 11.5%, but also the outage probability of the network decreases 23.6% by using the IHM.

本文首先在地铁光保真（LiFi）网络中建立接入点（AP）的信息价值模型（IVM）和相关的智能切换模型（IHM）。首先，基于统计方法，通过分析地铁乘客的行为特征创建IVM。通过IVM，可以成功预测用户访问每个AP的概率。接下来，基于这个概率，通过深入分析把手的交替阻塞问题，创建了 IHM。通过IHM，系统可以自适应迟滞状态。这不仅有效地避免了乒乓切换，而且降低了中断概率。最后通过仿真验证了上述IVM和IHM的有效性。仿真结果表明，IVM预测用户目标AP的准确率可达92%。特别是，与使用标准切换模型相比，使用IHM不仅网络吞吐量提高了11.5%，而且网络中断概率降低了23.6%。