Network Optimisation for Robotic Aerial Base Stations

Student thesis: Doctoral ThesisDoctor of Philosophy


One attractive application of unmanned aerial vehicles (UAVs) is to provide wireless coverage when acting as aerial base stations (ABSs). Compared to terrestrial small cells, ABSs have the benefit of flexible deployment, controllable mobility, and dominant line-of-sight channels, so they are expected to play a significant role in next-generation cellular networks. However, introducing this novel non-terrestrial communication device would also bring new challenges, such as requiring different evaluation criteria and being restricted by unexpected resource constraints. With this in mind, this thesis mainly focuses on the network optimisation problems of ABS-assisted networks.

Specifically, we first investigate two contradictory metrics, i.e., the information freshness and energy consumption, when an ABS is employed to collect data from ground terminals. A novel multi-return-allowed serving mode is proposed to explore the Pareto optimal trade-off between these two metrics. Secondly, to overcome the functional endurance issue of conventional ABSs, we propose a novel prototype named robotic aerial base stations (RABSs) with grasping capabilities, which can attach autonomously in lampposts or land on other tall urban landforms to serve as small cells with prolonged endurance. By employing this novel ABS prototype, we first study the optimal deployment and operation strategy for RABSs when the mobile traffic demand shows heterogeneity in both spatial and temporal domains. Afterwards, to further explore the use of RABSs in the upcoming 6G era, we investigate two novel application scenarios, that is, an RABS-assisted integrated sensing and communication (ISAC) system and an RABS-aided millimetre-wave (mmWave) backhaul network.

The proposed scenarios are formulated as various non-convex problems. By analyzing their constructions, we propose a variety of algorithms to solve them in a reasonable time. A wide set of simulation results shows that the proposed novel prototypes and serving schemes have immense potential in future cellular networks.

Date of Award1 Apr 2024
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorVasilis Friderikos (Supervisor) & Yansha Deng (Supervisor)

Cite this