LEO Satellite-Aided Over-the-Air Computation for Unmanned Aerial Vehicle Swarm Sensing

To overcome the bandwidth limitation for wireless data aggregation from a massive number of smart sensors or devices, over-the-air computation (AirComp) has been developed for ultra-fast data fusion. In this letter, we consider Low Earth orbit (LEO) satellite-aided AirComp networks to collect sensing information from unmanned aerial vehicle (UAV) swarms. To avoid excessive overhead in the conventional scheme, we propose a low-complexity pre- and post-processing scheme, where each UAV can adjust its phase in a distributed manner. Because the location information (LI) of the satellite fusion center (SAT-FC) is critical, we also derive the mean squared error (MSE) expressions for both perfect and imperfect LI cases. Analysis and simulation results show that the MSE may be degraded by a factor of 10 or more, with the azimuth and elevation angle errors of 0.001(degrees) and 0.1% distance error, which emphasizes the significance of the accurate LI of the SAT-FC.