UAVs over mmWave/THz Cellular MEC Networks: A Comparative Study for Energy Efficiency

Abstract: Cellular networks equipped with mobile edge computing (MEC) servers can be beneficial for unmanned aerial vehicles (UAVs) with limited onboard computation power and battery life-time. In this paper, we compare energy consumption of a UAV connected to cellular MEC servers in various possible scenarios such as onboard/MEC processing or parallel computation. Using detailed 3GPP-based modelings, we provide quantitative understanding of the most energy efficient approach and its relation with communication technologies, computation factors, and mobility parameters. Our findings show that, across the different frequencies from sub-6GHz to THz bands, the mmWave cellular MEC network is more energy efficient than UAV's onborad processing for a broader range of network densities. Secondly, while the UAV propulsion power consumption is non-decreasing function of velocity, the UAV movement cost yet can be optimized to further provide remarkable energy savings as compared to hovering. Finally, our results show that the most energy efficient approach can be obtained if mobility of UAV is combined with efficient parallel onboard-MEC processing.