Multiantenna UAV-Assisted Hybrid FSO/RF Data Collection for IoT: Optimal Design for Fairness

Wireless sensors are often deployed in hard-to-reach locations for data collection. Unmanned aerial vehicles (UAVs) can easily fly over such locations and so it is a promising solution to collect data from remote sensors. This article considers a UAV-assisted hybrid free space optical (FSO)/radio frequency (RF) data collection network for Internet of Things applications. In this network, multiple remote sensors transfer information to a UAV with multiple antennas through RF links using time division multiple access (TDMA). The UAV uses decode and forward protocol to send information to a base station (BS) via FSO link. Successive interference cancellation is employed to decode the information at the BS. Multiagent deep reinforcement learning is modified and applied to obtain a near-optimal scheme, for transmit power allocation, to minimize the maximum outage probability of decoding the information from the sensors under dynamic weather condition/UAV state. Numerical results are presented to illustrate the system design tradeoffs. Furthermore, the validity and superiority of our proposed approach is verified by comparing it with exhaustive search and differential evolution algorithms.