Predictive Communication for Unmanned Aerial Vehicle (UAV) Networks

Abstract

Unmanned Aerial Vehicles (UAVs), commonly known as drones, is an emerging technology with a huge potential to transform our lives into smart and connected communities by providing efficient infrastructure solutions for a wide range of military, industrial and commercial applications. UAVs empower innovative solutions that save lives and save the planet. However, there are several challenges and issues that limit the applicability of drones for many applications, not only due to UAVs' intrinsic and technological limitations, but also due to design and networking constraints. UAV networks composed of freely flying nodes create highly dynamic environments, where conventional networking protocols, which rely on stationary network contact graphs, fail to perform efficiently. Also, the efficiency of the networking protocols in terms of the incurred energy cost and the transmission delay can dramatically fall due to the networks failure in perceiving and predicting the environment. As a potential solution, Artificial Intelligence (AI) can revolutionize current networking methodologies by integrating computational intelligence into UAV networking solutions.

The aim of this project is to investigate the state of art of communication, computation and scheduling methods for UAV networking and propose novel solutions to solve the current issues and drawbacks. By using the prediction power of Machine Learning (ML) algorithms, we aim to better perceive the network topology, channel status, traffic distribution and resource availability to improve service provisioning. First, we propose three AI-enabled routing protocols for UAV networks that act based on learning from past history and anticipating future network states, yielding high performance for a variety of network scenarios and applications. Next, we suggest a predictive optimized compression policy for energy-efficient networking by avoiding excessive information exchange in dynamic scenarios. Last, we present an optimal sampling technique that reduces the interference and the overall energy consumption by a timely transmission of fresh data packets with considerable information content for Internet of Things (IoT). In summary, we believe that the offered solutions can revolutionize the current methods and pave the road to use UAVs in various IoT applications to benefit the society and the global economy.