|
You are at: Wagner
Home > Technologies > Data Fusion Tracking > Global Correlation Engine
Global Correlation Engine (GCE)
Global Correlation Engine (GCE)-for Naval Surface Warfare
Center-Dahlgren Division.
Daniel H.
Wagner Associates developed an information system for the processing
of localization and identification data from ESM/ECM, acoustic,
and other sources. GCE combines the best features of the MATCH
multi-hypothesis tracking and correlation system and the
Monte Carlo Computer Assisted Search (CAS) Search and Surveillance
Planning System (SSPS), but GCE is much more powerful than either
of these two systems. Using the GCE approach to processing all
of the available data concerning targets of interest will allow
the user to extract the maximum amount of information concerning
target location from the complex "sea" of available
data. The MATCH correlation algorithm is used to process large
volumes of data containing localization and attribute information
using multiple hypothesis methods. The SSPS Monte Carlo tracker is used to
compute contact association scores using non-Gaussian techniques which allow
for the accurate processing of non-Gaussian detection data, "negative
information" from unsuccessful search and the presence of land, and
information concerning probable target tactics.
| The resulting picture is
composed both of symbolic target locations from the standard Kalman filter
tracker and "probability maps" from the Monte Carlo process, giving the
operator a much more faithful picture of the combat picture and related
uncertainties. |
 |
GCE accepts reports from onboard sensors and from any available
ashore or afloat links. It processes positive information automatically,
and also processes any available negative and motion information
automatically.
In the Near-Real-Time Data Fusion (NRTDF)
project we produced a data fusion system using a new high-speed
data fusion architecture and massively parallel processing, to process hundreds or thousands of contacts per second in near-real-time
using multiple-hypothesis and non-Gaussian methods. The algorithms
in NRTDF are based on the ones developed for GCE, with the addition
of special pre- and post-processing modules.
|
