eXiT (Control Engineering and Intelligent Systems group)

Presentation

The eXiT group is an interdisciplinary research group (code GRCT 41) of the Institute of Informatics and Applications of the University of Girona involved in national and international research and transfer projects.

The main research activity of the group is focused on the application of artificial intelligence principles (data mining and knowledge discovery, qualitative reasoning, case based reasoning, auctions, etc.) in process monitoring and scheduling. Recently this activity has been extended to medical domain to analyse obesity and cancer patterns from historical data.

The members of the group participate in academic activities of the Master in Industrial Computing and Automatic Control and lead doctoral thesis in the PhD Program in Technology at the same University.

If you want to get in contact or visit us please see our contact information page.

Research Lines

The members of the group are active in the following research lines of the PhD. Program in Technology of the UdG:

Control Engineering and Intelligent Systems. Sub-lines:
- Model based fault detection and diagnostics.
- Multivariate statistical process control.
- Agent-based scheduling distributed systems.
- Monitoring power quality in electrical distribution systems.

Modelling, Control and Data Mining in Biomedicine. Sub-lines:
- Data mining for diagnosis and prognosis. Applications in cancer, obesity and diabetes.

Description

The activity of the group in the line of Control Engineering and Intelligent Systems is mainly devoted to study and develop new methods to cope with actual and future necessities of process monitoring and supervision. Main activity is focused on the efficient exploitation of data based on data mining and knowledge discovery principles:

Multivariate Statistical Process Control: Supervisory strategies for large number of variables. Dimensionality reduction and the development of fault detection and diagnosis mechanisms based on the process characterisation made from previous registers.
Case Based Diagnosis: Adaptation of Case Based Reasoning (CBR) methodology into the domain of dynamic process with emphasis on diagnosis. Definition of similarity criteria between signals, reuse of past experiences and definition of maintenance polices are in the scope of this line.
Qualitative representation of trends: Definition, formalization and development of numeric to qualitative interfaces. Representation of signals as sequences of episodes necessary to deal with knowledge based approaches.

Hybridation of techniques (computational and statistical) based on data mining principles for an efficient exploitation of historical data. This research is clearly domain dependent. Nowadays, the activity has been developed in during the last years in following domains:

Power Quality Monitoring: Automatic diagnosis of sags and fault location in distribution networks.
Batch Process Monitoring: Statistical monitoring of waste water treatment plants with Sequential Batch Reactors.
Off line Monitoring of ARIANE flights: Identification of abnormal flights from the analysis of sensor registers during the flights.
Blast Furnaces: On line monitoring of pressures to predict.
Moulding: Off line assessment of moulds (ageing and abnormal behaviours) based on injection pressures and temperatures.

Extension of this activity in the medical domain (Data Mining in medical Diagnosis and Prognosis research line) has been focused on the use of data mining principles to analyse historical clinical data of patients with obesity and cancer dysfunctions.

The activity in the research line Agent Based Distributed Scheduling Systems is focused on developing market-based mechanisms (auctions) for distributing shared resources and tasks among a set of agents. The application of resource/task allocation is very wide, ranging from distributing memory among processes in an operating system, to assigning providers to consumers, or selecting which is the most adequate agent to perform a task, amongst many other applications. Aspects as robustness and fairness are being now studied. Robustness is needed to assure the validity of a given solution front variations of the original scenario, whereas fairness is a condition necessary to guarantee the participation of the agents in terms of possibility to access to the resources they need.

Spin off

NEWRONIA is a Technologic Based Enterprise founded in 2011 as a result of the activity in the eXiT research group of Universitat de Girona to solve optimisation problems

Recogition /Awards

2009
eXiT is part of the Automation Engineering and Decision Support Systems (AEDS) research group, awarded with a consolidated distinction (2009 SGR 523) for the 2009-2012 period in the Consolidated Research Group (SGR) project of the Generalitat de Catalunya. The members o the group has renewed this distinction every four years since 1995.

Victor Muñoz was awarded with the ACIA Second Prize to The Best Master Thesis in Artificial Intelligence, offered by the ACIA (Catalan Association for the Artificial Intelligence) during the CCIA 2009 conference.

Xavier Berjaga was awarded with the ACIA Second Prize to The Best Master Thesis in Artificial Intelligence, offered by the ACIA (Catalan Association for the Artificial Intelligence) during the CCIA 2009 conference.
2008
Javier Murillo was awarded with the ACIA Prize to The Best Master Thesis in Artificial Intelligence offered by the CCIA (Catalan Association for the Artificial Intelligence) during the CCIA 2008 conference.

Javier Murillo, Víctor Muñoz, Beatriz López and Dídac Busquets were awarded with the Best Paper at the Sixth German Conference on Multiagent Systems Technologies.

Javier Murillo was awarded with the Best Student Paper at the Fourth Biennial Meeting: International Congress on Environmental Modelling and Software.

Javier Murillo and Víctor Muñoz were the 1^st classified in the 2008 International ART (Agent Reputation and Trust) Testbed Competition. Estoril, Portugal. May 12-16.
2007
Javier Murillo was awarded with the ACIA Second Prize to The Best Degree Thesis in Artificial Intelligence, offered by the ACIA (Catalan Association for the Artificial Intelligence) during the CCIA 2007 conference.

Javier Murillo and Víctor Muñoz were the 1^st classified in the Second Spanish ART (Agent Reputation and Trust) Testbed Competition. Valencia, Spain. March 26-27.
2006
Victor Muñoz was awarded with the ACIA Prize to The Best Degree Thesis in Artificial Intelligence, offered by the ACIA (Catalan Association for the Artificial Intelligence) during the CCIA 2006 conference.