Semantic Decision Support

The past years, there has been an increased use of data from sensors on the one hand and external systems on the other hand. Data, which can be used to enhance decisions made by stakeholders in different domains. Examples of such domains are healthcare, both home care and hospital/institutional care, the railway domain, smart cities and smart office/building environments. In order to extract information and knowledge from this multitude of data, semantics techniques are applied.

Our main research objective is to design algorithms and techniques that allow to consolidate all the available data and knowledge in order to offer context-aware and personalized decision support and actionable insights to the end-users.

Specifically, our research efforts focus on the following topics:

  •  Design of semantic knowledge models for structuring data originating from heterogeneous data sources
  • Design of Semantic decision support services and workflows, able to reason on the available data and alert or notify the appropriate stakeholder
  • Algorithms and techniques for semantic anomaly detection, root cause analysis and derivations of required actions
  • Algorithms for context- & social-aware task scheduling for robotic services
  • Particular focus and experience with designing services for intelligent transport, improving and supporting continuous homecare, realizing intelligent hospital information services and smart manufacturing.
  • Dynamic and semantic dashboards for decision support and root cause analysis

Staff

Filip De Turck, Femke De Backere, Femke Ongenae, Stijn Verstichel, Erik Mannens, Sofie Van Hoecke.

Researchers

Pieter Bonte, Gilles Vandewiele, Alexander Dejonghe, Jeroen Schaballi, Dörthe Arndt, Ben De Meester, Joachim Van Herwegen, Olivier Janssens.

Projects

  • ICON – OCareCloudS: Organizing Care through trusted Cloudy-like Services
  • ICON – FallRisk: Social-aware and context-aware multi-sensor fall detection system
  • ICON – Trapist: Train Passenger Interfaces for Smart Travel
  • ICON – WONDER: interventions for Wandering and Other behavioral disturbaNces of persons with DEmentia in nursing homes by personalized Robot interactions
  • IWT VIS SMARTPRO:Smart Textiles and wearable intelligence
  • Bilateral Ghent University Hospital on Advanced Alerting System

Key publications

Organizing home care, by fusing data from sensors and systems and information from both formal and informal caregivers, using intelligent, semantic decision support services.
Organizing home care, by fusing data from sensors and systems and information from both formal and informal caregivers, using intelligent, semantic decision support services.

 

The usage of semantic technologies to 1) consolidate all collected sensor data in a nursing home with the available background knowledge about the elderly and his/her environment, b) derive whether this person is currently exhibiting a behavioral disturbance, e.g., wandering, yelling, …, and c) decide which staff and/or robot intervention should be planned in order to resolve this behavioral disturbance and make the elderly feel at ease.
The usage of semantic technologies to 1) consolidate all collected sensor data in a nursing home with the available background knowledge about the elderly and his/her environment, b) derive whether this person is currently exhibiting a behavioral disturbance, e.g., wandering, yelling, …, and c) decide which staff and/or robot intervention should be planned in order to resolve this behavioral disturbance and make the elderly feel at ease.

 

Semantic alerting platform for Ghent University Hospital, allowing staff to define alerts on a semantic level. These alerts are then constantly checked on the huge amount of data collected about patients, such that the medical staff receive the appropriate notifications in a timely fashion.
Semantic alerting platform for Ghent University Hospital, allowing staff to define alerts on a semantic level. These alerts are then constantly checked on the huge amount of data collected about patients, such that the medical staff receive the appropriate notifications in a timely fashion.