Data driven economic operation of cyber-physical systems
- Principal Investigator
- David Muñoz de la Peña Sequedo / Daniel Limón Marruedo
- Funded by
- Plan Estatal 2013-2016 Retos – Proyectos I+D+i
- Partners
- Agencia Estatal Consejo Superior de Investigaciones Científicas
- Universidad de Huelva
Aligned with the H2020 programme, the document Plan Estatal de Investigación Científica y Técnica y de Innovación 2013-2016 (PEICTI) presents as main objectives the excellence in science, industrial leadership and tackling societal challenges in order to achieve a smart, sustainable and inclusive economic growth in Spain by coupling research and innovation. Information and communication technologies (ICT) is one of the areas of key industrial competences determining Europe’s global competitiveness and innovation across a broad range of private and public markets and sectors. The potential and capabilities of modern ICT systems are still growing exponentially fueled by the progress in electronics, microsystems, networking and data processing. ICT developments provide major opportunities for Spain and Europe to develop the next generation of open platforms for the management of increasingly complex cyber-physical systems (CPS) consisting of highly distributed and connected digital technologies that are embedded in a multitude of increasingly autonomous physical systems with various dynamics and satisfying multiple critical constraints including safety, security. power efficiency, high performance, size and cost.
This project aims to provide novel management systems for several CPS problems proposed by leading companies related to the 5th challenge in PEICTI “Action on climate change and efficient use of resources and raw materials” as well as the challenge on ICT application and solutions in the 7th challenge “Digital economy and society”. In particular to drinking water distribution networks and wastewater systems; HVAC systems in smart buildings and renewable energy generation plants will be studied.CPS are often characterized by very complex dynamics for which appropriate models are not always available, but because of the developments in sensing and communication techniques, largeamounts of data of the behavior can be obtained from historical operation data bases and online monitoring. Over the last years there has been a host of applications in different fields of novel data processing techniques such as government, internationaldevelopment, healthcare, media and advertisement. Standard automation techniques use historic data mainly for design purposes. The development of modeling, prediction, data reconciliation, fault detection and control techniques that take into account online in the decision making not only the knowledge of the system, but also the vast amount of data available, can potentially lead to groundbreaking advances.
In this project, new optimization based methods and tools that will profit from merging model and data information will be developed to ensure adaptability, scalability, complexity management, security and safety of CPS. Data driven methodologies based on the application of regularization approaches, pattern recognition and kernel methods will be studied. These methodologies will enable more precise state estimation and dynamic prediction algorithms as well as online-monitoring algorithms that will be able to identify non-desired emergent behaviors and faults. Based on this improved information, novel decision making algorithms for scheduling, control, fault tolerance and reconfiguration will be developed. The availability of large amounts of historic data will allow us to develop validation, performance assessment methodologies and address the human-in-the-loop issue through ad-hoc training.
