The project addresses the issue of providing advanced multimedia services by green Unmanned Aerial Vehicles (UAVs) networks.
Differently from the research lines described in the related works in the literature, in this project we address UAV networks trajectory planning strategies that are constrained by two key aspects: the energy consumed by the drone fleet on one side and the Quality of Experience characterizing the service offered to the users on the other side. As far as we know this is the first attempt to plan UAV paths jointly accounting for the energetic cost and the quality of the service provided by the drones as flying base stations. Let us point out that the project topic is highly stimulating from different points of view, and it opens virtually unlimited problems, since a variety of performance metrics can be considered as well as a large set of optimization parameters and, last but not least, different prior information. From a methodological point of view, the project will address theoretical and algorithmic developments as well as applications to real dataset. Furthermore, the activity is deeply rooted into national and international collaborations with US laboratories, and it will exploit of cutting edge technologies under different respect, such as i) drone radio interface modelling and programming, ii) deep learning based flight planning algorithms, iii) green/ brown energy availability models to cite a few, iv) UAV cognitive radio capabilities.
Given the technical challenges of the QoE-aware green UAVs flight planning problem, the expertise and the well-established cooperation of the project participant, and the promising outcomes of preliminary research already published by the project components, we expect the project to achieve timely, relevant, high impact research results.
The proposed project is inherently innovative with respect to the existing scientific literature for several main reasons.
Firstly, application of UAVs fleet to communication problems is relatively recent, and it has been boosted by the potential flexibility envisaged by 5G networks. Secondly, the project tackles with the problem of UAVs path planning under two different point of views, namely green mobile networking and QoE awareness. In this sense, it extends to the UAVs framework the holistic approach to multimedia networking problems that has been fruitfully applied in previous research by the components of the project team. Thirdly, the availability of extended video traffic traces databases inherited by previous projects and the extension of real data collection to different cellular environments boosts the interested of the developed methods.
From a methodological point of view, the work is deeply rooted into national and international collaborations with US laboratories, and it will exploit different cutting edge technologies, such as i)drone radio interface modelling and programming, ii)deep learning based flight planning algorithms, iii)green/ brown energy availability models to cite a few, iv)UAV cognitive radio capabilities.
From a timeline point-of view, stemming on the promising preliminar results obtained by the project participants on a single drone case study [13], we envisage the following project stages:
Stage 1: Extend the single-drone, QoE aware, energy constrained trajectory learning to the multi-drone case [13],[14], and evaluate the impact of several design choises (e.g. charging points locations) on the overall performances when an UAVs fleet is selected.
Stage 2: Generalize the theoretical formulation of the QoE- aware, energy constrained optimal flight planning to include different system priors, and develop greedy algorithms (including machine learning and deep learning ones).
Stage 3: Apply the identified algorithms to real data resulting from video traffic traces and urban/suburban coverage measurements, and implement the relevant functions of the algorithms on FPGA prototypes suitable for being drone mounted.
Given the technical challenges of the green, QoE aware UAVs flight planning problem, the differentiated expertise and the well-established cooperation of the project participant, as well as the promising outcomes of preliminary research already published in [13],[14], we expect the project to achieve timely, relevant, high impact research results.
[13] S. Colonnese, A. Carlesimo, L. Brigato, F. Cuomo, "QoE-aware UAV flight path design for mobile video streaming in HetNet" IEEE SAM 2018, Sheffield, July 2018.
[14] Ludovico Ferranti, Francesca Cuomo, Stefania Colonnese, Tommaso Melodia, "Drone Cellular Networks: Enhancing the Quality of Experience of Video Streaming Applications", Ad Hoc Networks, Available online 15 June 2018.