Eindhoven University of Technology (https://www.tue.nl/en/
) is one of Europe's top technological universities, situated at the heart of a most innovative high-tech region. Thanks to a wealth of collaborations with industry and academic institutes, our research has real-world impact. In 2015, TU/e was ranked 106th in the Times Higher Educational World University ranking and 49th in the Shanghai ARWU ranking (engineering). TU/e has around 3,000 employees and 2,300 PhD students (half of which international, representing about 70 nationalities).
The candidate will work at the Department of Electrical Engineering (https://www.tue.nl/en/university/departments/electrical-engineering/
). Within this department, research and education is done in domains of Telecommunication, Care and Cure, and Smart energy systems. The interfaculty Institute for Photonic Integration performs research in the area of broadband telecommunication, by investigating the potential of optical technologies. As a key member of IPI, the Electro-Optical Communication Systems (ECO) group focuses its research on optical communication system techniques, ranging from systems for ultra-high capacity long reach transmission (encompassing single-mode, multi-mode and multi-core fiber systems), ultra-fast (all-) optical packet switching nodes, high-density intra-data center networks, to multi-service flexible access and in-building networks (including radio-over-fibre and optical wireless communication). ECO participates in several national and international projects. Project description
The PhD project is in collaboration with KPN as part of the strategic co-operation between KPN and TU/e, TU/e-KPN flagship. The PhD project aims at investigating and demonstrating an AI-based automatic control of next generation photonics networks to support heterogeneous traffic growth from Mobile Access Network (5G and beyond) connected to Datacenters and edge computing nodes. The massive 5G cell deployments, datacenters, and edge computing nodes as well as novel applications with different requirements in terms of connectivity, latency and reliability is forcing operators such as KPN to deploy a large amount of networks elements and IT resources. Control and manage of such massive network elements and IT resources deployed in photonic networks becomes crucial to optimize and adapt those massive resources and match the heterogenous applications with different requirements as well as minimize their energy consumption. This will demand for novel low latency and scalable control systems. Autonomous operations and zero-touch networking are expected to enable a cost-effective and sustainable rapid photonic network growth. Zero-touch photonic networking is still in its early stages and requires the development of autonomous cognitive networking that entails closed-control-loop implementation collecting, analyzing, and making decisions on the photonic networks. AI-assisted decision making algorithms are growing importance for high-capacity photonic systems as the technique of choice to solve complex NP hard problems. The objective of this project is the investigation of novel network control systems exploiting AI-assisted network and IT resources optimization and allocation algorithms. The control plane will extensively rely on physical layer abstraction and modelling, and pervasive telemetry data collection to feed AI algorithms in order to implement and demonstrate the zero-touch networking paradigm.Tasks
The aim of the project is to develop and build an AI-assisted autonomous control system to support next generation real-time zero-touch photonic networks. To achieve this aim, specific procedures have to be designed to enable self-management, self-diagnosis, and self-optimization through monitoring the status of the photonic switches and transmission system as well the network traffic. All these procedures will be supported by new AI-empowered algorithms performing resource allocation and AI and telemetry-based monitoring providing predictive capabilities for innovative autonomous zero touch networks. The PhD candidate will design, develop, and experimentally assess an extensive physical layer abstraction and modelling of the optical network, including transmission system, photonic cross-connect switch, and IT resources. This includes the design and implementation of physical opto-electronic interfaces between the control plane and data plane in order to realize a pervasive telemetry data traffic collection from the physical optical network to feed the AI algorithms. Then the PhD candidate will develop novel AI-based algorithms for intelligent decision making embedding the developed model of the optical network as well as novel AI-based traffic prediction models to implement a proactive resource allocation and enhance the performance and resource utilization of the network. Based on the developed AI-algorithms, the PhD candidate will design a framework for an AI-assisted autonomous and dynamic network supporting real-time operations and zero-touch network, including the development of autonomous cognitive networking that entails closed-control-loop implementation collecting, analysing, making decisions, and acting on the network devices. The PhD candidate will demonstrate and assess in the ECO laboratories the system operation and performance of the AI-assisted ultimate zero-touch networking control and telemetry systems for the resource optimization and allocation of the optical network interconnecting datacenters and multiple edge computing nodes to mobile access networks (5G and beyond). The PhD candidate will contribute to the TU/e efforts in establishing collaboration with the other researchers in the project and contribute to the related project reporting, scientific publication and dissemination activities.