Aznar Fidel, Pujol Mar, Rizo Ramón, Rizo Carlos
Department of Computer Science and Artificial Intelligence, University of Alicante, San Vicente del Raspeig, Alicante, Spain.
Department of Architectural Constructions. University of Alicante, San Vicente del Raspeig, Alicante, Spain.
PLoS One. 2018 Jan 25;13(1):e0190692. doi: 10.1371/journal.pone.0190692. eCollection 2018.
In this work, a swarm behaviour for multi-rotor Unmanned Aerial Vehicles (UAVs) deployment will be presented. The main contribution of this behaviour is the use of a virtual device for quantitative sematectonic stigmergy providing more adaptable behaviours in complex environments. It is a fault tolerant highly robust behaviour that does not require prior information of the area to be covered, or to assume the existence of any kind of information signals (GPS, mobile communication networks …), taking into account the specific features of UAVs. This behaviour will be oriented towards emergency tasks. Their main goal will be to cover an area of the environment for later creating an ad-hoc communication network, that can be used to establish communications inside this zone. Although there are several papers on robotic deployment it is more difficult to find applications with UAV systems, mainly because of the existence of various problems that must be overcome including limitations in available sensory and on-board processing capabilities and low flight endurance. In addition, those behaviours designed for UAVs often have significant limitations on their ability to be used in real tasks, because they assume specific features, not easily applicable in a general way. Firstly, in this article the characteristics of the simulation environment will be presented. Secondly, a microscopic model for deployment and creation of ad-hoc networks, that implicitly includes stigmergy features, will be shown. Then, the overall swarm behaviour will be modeled, providing a macroscopic model of this behaviour. This model can accurately predict the number of agents needed to cover an area as well as the time required for the deployment process. An experimental analysis through simulation will be carried out in order to verify our models. In this analysis the influence of both the complexity of the environment and the stigmergy system will be discussed, given the data obtained in the simulation. In addition, the macroscopic and microscopic models will be compared verifying the number of predicted individuals for each state regarding the simulation.
在这项工作中，将展示一种用于多旋翼无人机（UAV）部署的群体行为。这种行为的主要贡献在于使用了一种虚拟设备进行定量半构造性符号互动，从而在复杂环境中提供更具适应性的行为。这是一种容错性高且鲁棒的行为，它不需要待覆盖区域的先验信息，也无需假定存在任何类型的信息信号（全球定位系统、移动通信网络等），同时考虑到了无人机的具体特性。这种行为将面向应急任务。其主要目标是覆盖环境中的一个区域，以便随后创建一个自组织通信网络，该网络可用于在该区域内建立通信。尽管有几篇关于机器人部署的论文，但更难找到无人机系统的应用，主要是因为存在各种必须克服的问题，包括可用传感和机载处理能力的限制以及飞行续航能力低。此外，那些为无人机设计的行为在实际任务中的使用能力往往有很大局限性，因为它们假定了特定特征，不容易以通用方式应用。首先，本文将介绍模拟环境的特征。其次，将展示一个用于自组织网络部署和创建的微观模型，该模型隐含地包含了符号互动特征。然后，将对整体群体行为进行建模，提供这种行为的宏观模型。该模型可以准确预测覆盖一个区域所需的智能体数量以及部署过程所需的时间。将通过模拟进行实验分析以验证我们的模型。在该分析中，将根据模拟中获得的数据讨论环境复杂性和符号互动系统的影响。此外，将比较宏观和微观模型，验证关于模拟的每个状态下预测个体的数量。
