Deep Learning-based LAP Deployment and Aerial Infrastructure Sharing in 6G

Sixth Generation (6G) considers Low Altitude Platform (LAP) based aerial cells for improving coverage and capacity in cellular networks. The available literature on LAP based aerial cells has addressed the altitude and trajectory optimization for LAPs. However, aerial cell sharing amongst operators for capacity enhancement and resource utilization is not studied enough. To the best of our knowledge, for the first time, this paper proposes, (a) four architectures for aerial infrastructure sharing, (b) a novel element, i.e., Aerial Asset Manager (AAM) for deployment and sharing of LAPs by operators from one of the above four architectures. AAM utilizes a Long Short-Term Memory (LSTM) based deep learning model for traffic forecasting. The forecast helps to plan the deployment and opportunistic sharing of LAPs amongst operators. The analytical model and simulations results reveal the forecasting accuracy of 96% for data traffic. AAM improves the capacity for operators, maximizes the resource utilization of the LAPs, i.e., the bandwidth utilization increases by 49%, and minimizes the number of LAPs, i.e., the fleet size reduces by 45% in comparison to the available literatures.