Infrastructure for Information Fusion (StrucFus)

The purpose of this project is to consider infrastructure issues for the software engineering of and operation of applications in information fusion. In particular, real-time processing as well as distributed real-time databases are of major interest. The reason is that timely processing of information is a major concern in information fusion, since we deal with support for automatic or semi-automatic decision-making where the decision makers need stringent response times to do their work. This timely processing is complex since we must manage the uncertainty involved in the acquired data in information fusion while handling the constraints of existing resources (e.g., network bandwidth). Further, information fusion systems often have requirements on dependability (e.g., reliability and availability requirements). Typically, in many information fusion applications, critical data must be highly available, and since these applications are used in unsafe situations (e.g., post-disaster management), we must replicate data to retain it in case the system crashes.

 

The aim of the collaborative project

The aim of the project is to bring together high quality researchers with a basis in infrastructural issues (real-time systems, real-time databases, operating systems etc.) with a specific interest in information fusion or related fields in order to define the infrastructure for information fusion systems. The goals are primarily to (i) investigate the key problems, key solutions, and key services of relevant research areas to check if these are required or desirable in an infrastructure for information fusion systems; (ii) define the components and their interactions (preferably by refining an existing infrastructure for fusion systems), and (iii) compare and contrast this architecture to existing relevant architectures. By necessity, an infrastructure for information fusion systems must be configurable since many disparate applications are found in information fusion as well as consist of different techniques, algorithms, and tools.
 
 


Uppdaterad: 2010-05-10
Sidansvarig: Marcus Brohede