Researchgroups in The Netherlands
The Distributed Hybrid Fusion Group (DHFG) at the D-CIS lab is focusing on the development of sound approaches to modeling and processing which support distributed information fusion and collaborative reasoning with humans in the loop.
The DHFG’s activities include basic research on theoretical aspects of modeling and inference with Bayesian networks as well as the development of service oriented architectures and tools that facilitate implementation of real world fusion systems.
The basic research on Bayesian networks is addressing the grounding problem in multiple, complementary ways. In particular, we have been investigating
(i) the impact of modeling deviations and inherently robust fusion systems
(ii) sound modularization of Bayesian models and inference processes supporting decentralized fusion and
(iii) the use of structure learning techniques for the discovery of modeling faults in distributed Bayesian systems.
The theoretical results are being implemented in software components which support construction of robust self organizing distributed fusion systems
Currently, the developed techniques are being evaluated in collaboration with the DCMR Milieudienst Rijnmond in a real world prototype supporting detection and localization of chemical leaks.
Moreover, recently multiple studies on the use of the theoretical and technical results in other domains, such as crisis management in general, security, etc. have been carried out.
The DHFG is closely collaborating with multiple academic and user partners throughout Europe.
Dr K. Nieuwenhuis
Dr G. Pavlin