Mobility-Aware Pre-caching Based on Unsupervised Deep Generative Model for Small Cell Networks
Zhen Yao, Shuangwu Chen, Jian Yang, Yifeng Liu, Haiyong Xie · 2019
Due to the short coverage of small-cell base stations (SBS), mobile device is more likely to move away from its covered area, and thus has to continually switch between SBSs and reestablish the connections with the remote servers. Aiming for improving the delivery efficiency of ever-growing mobile data traffic, we conceive proactive caching over dense small-cell network (DSCN). Without any prior knowledge, it is challenging to Figure out users movements and determine content placement in SBSs. Motivated by the recent advances in deep learning, we first propose a model-free mobility prediction approach based on the conditional variational autoencoder (CVAE) in this paper. Our approach is able to infer the latent information about users habits, which are considered to be closely related to their movements, from historical trajectory. By testing on real-world GPS trajectories, our approach achieves a prediction accuracy of about 80%. Then based on the movement prediction, we formulate an optimization problem to maximize the local cache utility for DSCN. The experimental results show that our mobility-aware pre-caching strategy can support a seamless mobility handover with a lower delay and a higher data rate.