ABSTRACT
The features of multiple and collaborative robots have been explored both in industry and academia to address a number of problems in several application domains and several open research problems. Swarms composed of unmanned aerial vehicles (UAVs) can be seen as important examples of such multiple robotics systems, which are being able, in an increasingly more powerful away, to sense the environment, process the collected data and communicate with other devices using one or more wireless media access networks. The ability to have UAVs communicating among each other also enhances the possibilities of having them operate autonomously or at least with less human dependency. UAVs may be used for detecting and applying substances in the agricultural fields, combating fire in vast forest areas or even searching for victims in disaster scenarios, among many other applications. Aiming to have an autonomous operation, the network management of flying ad hoc networks (FANETs) that supports the UAV swarm operation can be performed in a centralized, decentralized or distributed way. In a centralized approach, the network has a unique controller, which is in charge of controlling many aspects of the network, such as its topology and routing policy. In a decentralized approach, the control is hierarchically subdivided into many subcontrollers. Thus, the main controller does not need to have an entire view of the network. In a distributed approach, the nodes act independently through the network, which means that they only have a view of its vicinity, and the network does not depend on a central node; thus, the network does not present a unique point of failure. For every type of network management and for every kind of collaborative application using UAVs, the system will depend on a drone-to-drone (D2D) communication, which can be really challenging due to a drone's high mobility nature. In light of this challenge, this chapter discusses alternatives found in the literature to address the problem, proposing a software and hardware architecture able to implement an efficient communication support for UAV swarm systems. The performance, features and limitations of this solution are discussed by analyzing results acquired with field experiments.
