Accession Number : ADA636858
Title : A Consistent Filter for Robust Decentralized Data Fusion
Descriptive Note : Technical rept.
Corporate Author : DEFENCE RESEARCH AND DEVELOPMENT CANADA VALCARTIER (QUEBEC)
Personal Author(s) : Benaskeur, Abder R ; Roy, Jean
Report Date : 29 Oct 2002
Pagination or Media Count : 73
Abstract : The Situation Analysis Support Systems (SASS) Group in the Decision Support Systems (DSS) Section at Defence Research & Development Canada (DRDC) - Valcartier is currently investigating advanced concepts for adaptation and integration of the data fusion and sensor management processes. These concepts could apply to any current Canadian military platform's sensor suite, as well as its possible future upgrades, to improve its performance against the predicted future threat. The reported work addresses the problem of automatically aggregating information from multiple data sources. Multiple Source Data Fusion (MSDF) is used to indicate the general approach for combining the sensor data into global tracks. The selection of the appropriate MSDF techniques depends on the underlying architecture. For the centralized scheme, the sources are known to be independent and the Kalman filter provides an optimal solution. Unfortunately, when the decentralized architecture is used the sources become correlated and the Kalman filter cannot be applied. The covariance intersection method has been proposed as a solution to the problem of the decentralized data fusion, but results in a decrease in performance. A new fusion algorithm, that avoids both the inconsistency of the Kalman filter and the lack of performance of the covanance intersection, is proposed. The superiority of the proposed approach is illustrated through the target's tracking problem.
Descriptors : *DATA FUSION , ALGORITHMS , CENTRALIZED , CORRELATION , COVARIANCE , DECENTRALIZATION , ELLIPSOIDS , KALMAN FILTERING , TRACKING
Subject Categories : Statistics and Probability
Distribution Statement : APPROVED FOR PUBLIC RELEASE