A probabilistic optimal UAV trajectory planning approach to minimise the uncertainty of traffic density estimations

© 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Traffic monitoring has been one of the major tools used for transportation operations and planning. The emergence of Unmanned Aerial Vehicles (UAVs) provides new capabilities for traffic management purposes. One of the most challenging tasks of UAVs is path planning, that aims to determine the most efficient route of a UAV from an initial point to a target point, with respect to a specific task to be optimised. In this work we focus on the efficient traffic density estimation. Towards this, we propose an online probabilistic UAV trajectory construction methodology, that aims to select the next moves from which the UAV should obtain measurements of the traffic density, while aiming to minimise the total uncertainty of the traffic density across all time-space points in the time-horizon under study. We incorporate the Gaussian Process (GP) model to accurately estimate traffic density of specific road segments in a multi-lane highway even when data points are sparse within the specified time-space region under study. We employ a decision-theoretic methodology to develop Bayesian optimal UAV trajectories aimed at minimizing traffic density uncertainty. Finally, a simulation study is performed to evaluate the proposed methodology that shows that it reduces the uncertainty of the traffic density up to 70% compared to a simple cyclical UAV trajectory.