Multi-Objective Optimization for Data Collection in UAV-Assisted Agricultural IoT

To enhance the observation and management of agricultural plants, it is critical to collect data from Internet of Things (IoT) devices in agriculture. However, the use of fixed ground-based stations (BSs) often results in inflexible deployment, high overhead costs, and increased vulnerability to damage from natural disasters, which can impede continuous data collection. To address these challenges, this work explores the use of Unmanned Aerial Vehicles (UAVs) as aerial BSs to gather data from IoT devices. First, we formulate a UAV-assisted data collection multi-objective optimization problem (UDCMOP) to efficiently collect agricultural data. Specifically, we aim to collaboratively optimize the hovering positions of UAV, visit sequence of UAV, speed of UAV, and the transmit power of devices, to simultaneously maximize the minimum transmit rate of devices, minimize the total energy consumption of devices and UAV. Second, the proposed UDCMOP is characterized as a non-convex mixed integer nonlinear optimization problem, containing both continuous and discrete variables, which presents considerable challenges in terms of solvability. Therefore, we solve it by proposing an improved multi-objective artificial hummingbird algorithm (IMOAHA) with several specific improvement factors, including the hybrid initialization operator, Cauchy mutation foraging operator, and the discrete mutation operator. Simulations are carried out to testify that the proposed IMOAHA can effectively improve the system performance in comparison to existing benchmarks. Additionally, to verify the effective working time of the UAV system, we investigate both random and uniform UAV deployment strategies and consider the impact of varying farm topology on the system model. Finally, practical implementation experiments using a Raspberry Pi confirm the feasibility and effectiveness of the proposed UAV-assisted communication system in real-world environment.