An adaptive UAV positioning model for sustainable smart transportation

Abstract

• Using optimization to dynamically adjust the drone height considering vehicular traffic and energy consumption. • Predict vehicular traffic patterns. • Developed an adaptive model that allows to adapt the position of the drones based on the predicted vehicular traffic. • Using Python, scilkit-learn library, real street traffic data, and NS-3, we showcase the effectiveness of the adaptive model in terms power usage, traffic prediction accuracy, and network coverage. Several research works are being considered to adopt the use of UAVs to support smart transportation systems due to their movement flexibility. In this article, a UAV-supported vehicular network solution is developed which considers both power and coverage limitations of UAVs to attain the vision of sustainable smart cities. Nodes communicate with each other through the 5G connection and ad-hoc links. The solution is solved for as a predictive optimization problem that determines the height of the UAV to dynamically change it to ensure the optimal communication coverage of vehicular nodes. Moreover, the solution considers UAV energy consumption constraints when setting the optimal height of the UAV. Additionally, the optimal distance between every two adjacent UAVs is also considered to avoid any coverage overlapping while protecting their connectivity. Extensive evaluations were considered in terms of both implementation and simulation to test the proposed model. Evaluation results show that the proposed solution can predict vehicle traffic patterns accurately to ensure proper adjustments of the UAV height. Moreover, network coverage is ensured for areas with and without fixed BS availability with the support of the self-positioning UAVs while adhering to QoS requirements.