Unmanned Aerial Vehicle Allocation and Deep Learning based Content Caching in Wireless Network

Data traffic is increasing with the increasing number of smart devices. Also, base stations in some regions are suddenly overloaded only for a certain period (i.e., an amusement park on holiday). Thus, to handle this issue, we need to deploy more base stations, small-cell base stations. But those are not economical solutions. Hence, in this paper, we utilized Unmanned Aerial Vehicles (UAVs) as temporary small-cell base-stations to solve the aforementioned problems. In this work, we proposed a cluster-based UAVs deployment scheme to reduce data traffic (such as video traffic) as well as service delays for the users and improve the coverage of base stations. First, we formed the user groups according to the distance of users with the help of the K-means clustering algorithm. Second, we find the optimal location to allocate the UAVs in each cluster. Third, we proposed a Long Short-Term Memory based caching scheme to cache popular contents on UAVs. Finally, the simulation results show that our proposed scheme outperforms than the other in terms of accessing delay and cache hit ratio.