In emerging wireless networks, the scalability of deploying drones presents an opportunity to design extensive aerial networks. These networks could effectively monitor large agricultural fields from the air and soil for food production with efficient resource utilization. On the one hand, unmanned aerial vehicles (UAVs) have gained interest in agricultural aerial inspection due to their ubiquity and observation scale. On the other hand, agricultural internet-of-thing devices, including buried soil sensors, have gained interest in improving natural resource efficiency in crop production. In this work, we investigate the natural interaction of these two phenomena, where UAVs can be leveraged as flying nodes to collect information from above-ground (AG) and underground (UG) nodes. Additionally, UAVs play a crucial role in bridging farmlands with the core network, facilitating the provision of essential network services. In such a scenario, the UAV can be easily imagined to communicate with AG nodes, UG nodes, neighboring UAVs, and the primary base station (BS) that connects the aerial network to the core network. Therefore, we consider several communication links in a UAV-X communication scenario in a multi-UAV network that is imagined to be deployed by a primary BS to monitor a large field, where X refers to the ground node, UG node, BS, or neighboring UAV. We also consider that the primary BS is equipped with a mobile edge computing server (MEC) that not only assists the aerial network but also serves the communication and computational needs of the respective ground users. Following that, four links can be defined: the Air-to-Air (A2A) link, the Air-to-Ground (A2G) link, the Air-to-Underground (A2UG) link, and the Ground-to-Ground (G2G) link, all of which are considered in this work to study various research problems.

First, we investigate the multi-antenna channels between two UAVs in terms of antenna correlation and system capacity in the A2A link scenario. The effect of 3D position on multi-antenna channel characteristics is investigated, and significant variation in the channel is observed in relation to the azimuth and elevation angles between the UAV nodes. Based on the findings, we propose an effective machine learning-based technique for estimating the direction of a transmitting node in an A2A link.

Second, we study a UAV-based full-duplex (FD) multi-user communication network in the A2G link scenario, where a UAV is deployed as a multiple-input--multiple-output (MIMO) FD BS to serve multiple FD users on the ground. A novel multi-objective resource allocation problem is designed and solved, which maximizes the sum uplink (UL) and downlink (DL) rates while optimizing the DL beamformer, beamwidth angle, 3D position of the UAV, and UL power of the FD users.

Third, we investigate path loss and fading characteristics between UAV and UG nodes using outdoor measurements, aiming to facilitate energy-efficient data collection to and from A2UG wireless links. A novel model is developed that estimates path loss with reduced errors across various UAV 3D positions than prior models. Accordingly, an energy-efficient aerial data collection strategy is designed.

Last, in the G2G link scenario, we consider a network in which a BS associated with an MEC server provides computing services to uplink user equipment (UUE) and downlink user equipment (DUE). By leveraging FD at the BS, we design a novel time-slotted computational task completion protocol that can efficiently use computation and communication resources in the network. In this setup, we jointly optimize the BS transmitter precoding vector, UUE uplink transmit power, MEC computational resources, and time-slotted computational task shares to minimize the sum weighted energy at the UUE and server while satisfying a completion-time threshold for each user's task.

Degree Date

Spring 2024

Document Type


Degree Name



Electrical and Computer Engineering


Joseph Camp

Number of Pages




Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License