Deep Learning based Efficient Trajectory Planning and Sharing for Aerial Cells in mMTC Networks

Terrestrial cellular networks can benefit immensely for both coverage and capacity extension by deploying Low Altitude Platform (LAP)-based aerial cells. Reduction of the capital expenditure (CAPEX) and operation expenditure (OPEX) is a key consideration by telecom operators and sharing of aerial cells can be a great enabler. The available literature on aerial cell sharing amongst operators is limited to the use case of capacity enhancement in enhanced mobile broadband (eMBB), while use cases of massive machine type communication (mMTC) are not analyzed. To the best of our knowledge, for the first time, this paper proposes a novel mMTC-enhanced-Aerial Asset Manager (mMTC-enhanced-AAM) algorithm for trajectory planning and sharing of a LAP-based aerial cell by multiple operators in an mMTC network. mMTC-enhanced-AAM utilizes a deep learning model for data traffic forecasting and using the available hovering time constraint in LAP-based aerial cells, presents a predicted epoch-based time-sliced sharing plan amongst operators. The presented model and simulation results reveal much lower error (less than 10%) in mMTC data traffic forecasting. mMTC-enhanced-AAM minimizes the fleet size for LAPs and improves the utilization of the aerial cells, where the hover time utilization improves by 78%, and the number of LAPs requirement by operators reduces by 34% than the available literature.