XDense: A Mesh Grid Sensor Network for Extreme Dense Sensing
Ref: CISTER-TR-190802 Publication Date: 24, Jan, 2020
XDense: A Mesh Grid Sensor Network for Extreme Dense SensingRef: CISTER-TR-190802 Publication Date: 24, Jan, 2020
We introduce XDense, a novel sensor network system for application scenarios that could benefit from densely physically deployed sensors. More specifically, XDense was conceived for cyber-physical systems (CPS) that require real-time dense sensing, for example, involving thousands of sensors per square meter in real-time. We motivate our work by presenting CPS application scenarios that could potentially benefit of dense sensor networks, which are currently limited by available technology. Out of the different application fields we discuss, we give special focus to avionics. More specifically, we focus on active flow control (AFC) on aircraft wing surfaces. We aim at providing means of sampling data with a high spatial and temporal granularity about the air flowing through an aircraft wing, so that active control over the aerodynamics of the wing is feasible. The XDense architecture consists of a wired 2D-mesh grid network that provides it distributed processing capabilities, that are used to enable real-time complex environmental data extraction in a distributed fashion. It resembles Networks-on-Chip (NoC) architecture, principles of operation and temporal behavior. The similarities and differences are discussed. We detail XDense’s node and network architecture, protocols, and principles of operation. We evaluate the performance of XDense in fluid dynamic application scenarios with extensive experiments on sensing and feature detection capabilities. We also tackle the issue of time predictability of XDense. We present a methodology that uses traffic shaping heuristics to guarantee bounded communication delays while fulfilling memory constraints. We evaluate the model for multiple network configurations and workloads, and present a comparative performance analysis in terms of link utilization, queue size and execution time. With the proposed traffic shaping heuristics, we endow XDense with the capabilities required for real-time applications. We also discuss the practical issues involved in implementing XDense and the steps for its experimental validation. A prototype node and a test-bed was implemented to validate our assumptions and to assess the performance capabilities.
PhD Thesis, FEUP.
Notes: Presidente do Juri Doutor José Alfredo Ribeiro da Silva Matos, Professor Catedrático da FEUP Vogais Doutor Leandro Soares Indrusiak, Professor do Department of Computer Science da The University of York, United Kingdom; Doutor Jean-Luc Scharbarg, Professor da IRIT-ENSEEIHT – l’École Nationale Supérieure d'Électrotechnique, d'Électronique, d'Informatique, d'Hydraulique et des Télécommunications, Toulouse; Doutor Eduardo Manuel Medicis Tovar, Professor Coordenador do Instituto Superior de Engenharia do Porto (Orientador); Doutor João Manuel Paiva Cardoso, Professor Catedrático do Departamento de Engenharia Informática da Faculdade de Engenharia da Universidade do Porto, Doutor Paulo José Lopes Machado Portugal, Professor Associado do Departamento de Engenharia Eletrotécnica e de Computadores da Faculdade de Engenharia da Universidade do Porto.
Record Date: 27, Aug, 2019