Application of machine learning in the internet of things that compose the cyber-physical ecosystem of smart cities

Abstract

This grant is an integral part of "Plan of institutional development in the area of digital transformation: advanced manufacturing and smart and sustainable cities (PDIP)" of the Technological Research Institute of the State of São Paulo already approved by FAPESP under the Process 2017/50343-2. The IPT PDIP, which runs from May 1, 2018 until April 30, 2021, aims to, among other objectives, enable the Center for Information Technology, Automation and Mobility - CIAM "in cyber-physical systems that allow the transfer of physical phenomena to a digital environment, deploying a platform to support intelligent solutions for cities and industry. "Cyber-Physical Systems can be defined as the fusion of physical elements, such as sensors and actuators, working side by side with cyber-physical elements such as software to monitor and initiate physical processes, as well as record and analyse previously stored data for decision making appropriate to the context. In addition, simultaneous and equally rapid progress in the field of communications and the Internet of Things allowed embedded systems to be equipped with the power of collective knowledge, gathered from the integration of information shared by other things, systems, individuals and authorities, as opposed to working in isolation. As a result, Cyber-Physical Systems, Citizens and users of this ecosystem can initiate actions, almost in real time, to make cities more efficient and reliable, allowing greater optimized interactions based on those feedbacks. The BPE will be held from April 8 to November 7, 2019 at the Technische Universität Berlin/TUB, based on the Network Information Theory Group, under the guidance of Dr. Renato Luís Garrido Cavalcante. Research and training activities will be carried out in machine learning techniques geared to the processing of data from the internet of the things that make up the Intelligent Cities Cyber-Physical Ecosystem. An important part of these activities will be held in the Network Information Theory Group, interacting with projects in the lines of research: 1) Machine Learning; 2) Sensor Networks; 3) Distributed Signal Processing; 4) Convex Analysis; 5) Wireless Communications.Within the PDIP, such techniques will be applied in simulation, control and optimization models of operations in real systems found in industrial and urban environments. The motivation in choosing the Technische Universität Berlin is not limited to its excellent academic reputation in the use of simulation models applied to cyber-physical systems, machine learning, sensor network and internet of things, but it extends to its portfolio of projects carried out in partnership with the Fraunhofer-Gesellschaft, in particular with the Fraunhofer Institute of Telecommunications, Heinrich Hertz Institute, HHI. Fraunhofer HHI has accumulated great experience in communication networks in both 5G technology, which will be the Internet of Things backbone, as well as industrial networks. The computational resources they use for simulation may be useful in IPT teaching activities and in industry 4.0 projects to be developed at CIAM. Finally, participation in seminars, writing of academic articles, participation in research projects, and visits to projects of other research institutions are also foreseen. (AU)