Department of Informatics

TU München - Fakultät für Informatik
Software & Systems Engineering Research Group

TUM
 
 

Project "CawarFlow"

The CawarFlow project constitutes an interdisciplinary project between the chair IV of the department of informatics of the Technische Universität München and the MITI group of the faculty of medicine of the Technische Universität München. It has an estimated duration of 2 years and is funded by the Deutsche Forschungsgemeinschaft.

The subject of CawarFlow is the usage of domain knowledge in form of workflow descriptions for enhancing the efficiency and accuracy of automatic adaptations performed by context adaptive software systems. The project results are evaluated by means of an attendant case study from the domain of minimally invasive medicine. We focus on providing computational support for a largely standardized surgery, namely the laparoscopic cholecystectomy (removal of gallbladder).

A main objective of this project is the reduction of Unwanted Behavior (UB) - a specific kind of malfunction, which is observable when deploying context adaptive systems in the real world. The integration of workflows, which reflect the knowledge concerning the interaction of domain specific processes and the operational environment, is supposed to enable context adaptive systems to better recognize and predict actual situations and associated needs.

Therefore a semantically founded workflow description is needed, which offers an appropriate level of granularity and variation for reflecting the precise course of the considered surgery. The workflow model is finally integrated into the context adaptive system, which infers on basis of this workflow and the available sensors within the operating room. This enables the system to monitor and observe the surgery progress and to adopt certain tasks like recommending next steps or alerting auxiliary personnel in critical situations. In consequence, several organizational and operational tasks are facilitated, while the patient safety is enhanced by anticipating future events and triggering the appropriate interventions.

© 1999-2009 - Software & Systems Engineering Research Group
Sitemap |  Contact
Last change: 2010-06-07 14:47:28