Trajectory Optimization of Cellular-Connected UAV for Information Collection and Transmission

TL;DR

This paper develops a trajectory optimization framework for cellular-connected UAVs to efficiently collect and transmit information from multiple ground targets, balancing mission time and communication constraints.

Contribution

It introduces a structured communication protocol and a graph-based model to efficiently optimize UAV trajectories for information collection and transmission.

Findings

Proposed algorithm reduces mission completion time.

Outperforms benchmark schemes in numerical simulations.

Achieves efficient information collection and transmission.

Abstract

In this paper, we consider a cellular-connected unmanned aerial vehicle (UAV) with an information collection and transmission mission for multiple ground targets. Specifically, the UAV is required to collect a fixed amount of information of each target by hovering at a pre-determined location (via e.g., photography/videography/sensing), and transmit all the collected information to the cellular network during its flight. We aim to jointly optimize the UAV's trajectory and the information collection order of the ground targets to minimize the mission completion time. The formulated problem is NP-hard due to the need of visiting the information collection locations for all targets; moreover, the UAV's trajectories over different time durations are coupled in non-convex constraints for ensuring information transmission completion. To handle this difficult problem, we first propose a structured communication protocol between the UAV and the cellular network, which decouples the UAV's trajectory designs in different time durations. Then, under the proposed protocol, we establish an equivalent graph-based model for the considered problem, and devise a low-complexity algorithm for finding an approximate solution by exploiting the problem structure and leveraging graph theory. Numerical results show that our proposed design achieves efficient information collection and transmission, and outperforms various benchmark schemes.