Decentralized Q-learning based Optimal Placement and Transmit Power Control in Multi-TUAV Networks
Optimizing the placement and transmit power control of unmanned aerial vehicle-base stations (UAV-BSs) is a key priority in 6G air-to-ground communication networks. A multi-UAV network brings many advantages, such as high line-of-sight (LoS) probability, three-dimensional (3D) connectivity, flexible mobility, and cost-effectiveness. However, it still has a problem of severe battery constraints. To overcome this battery problem, the concept of a tethered UAV (TUAV) is introduced, and it receives sufficient battery power from the ground through a tether. Therefore, we propose a decentralized Q-learning-based optimal placement and transmit power control algorithm (DQ-OPP) to maximize the individual data rate of each TUAV-BS. Through simulations, we show that the proposed DQ-OPP algorithm outperforms the conventional algorithms.
Paper
Full text
Decentralized Q-learning based Optimal Placement and Transmit Power Control in Multi-TUAV Networks
Semantic Scholar · Engineering · 2023
Abstract
Optimizing the placement and transmit power control of unmanned aerial vehicle-base stations (UAV-BSs) is a key priority in 6G air-to-ground communication networks. A multi-UAV network brings many advantages, such as high line-of-sight (LoS) probability, three-dimensional (3D) connectivity, flexible mobility, and cost-effectiveness. However, it still has a problem of severe battery constraints. To overcome this battery problem, the concept of a tethered UAV (TUAV) is introduced, and it receives sufficient battery power from the ground through a tether. Therefore, we propose a decentralized Q-learning-based optimal placement and transmit power control algorithm (DQ-OPP) to maximize the individual data rate of each TUAV-BS. Through simulations, we show that the proposed DQ-OPP algorithm outperforms the conventional algorithms.