|
You are at: Wagner
Home > Technologies > Data Fusion Tracking > Match
Multi-Hypothesis Correlation
(MATCH)
MATCH is a correlation and tracking algorithm, designed initially
for tracking ships using reports of the targets' ELINT and other
attributes along with their position and velocity. It accepts
data in the form of contact reports, and uses them to form estimates
of the number of targets present, their positions and velocities,
and their ELINT and other characteristics. For this purpose, MATCH
uses a collection of attribute models, which are probabilistic
models of the behavior of the targets under surveillance and of
the process by which reports are generated. The most important
of these models is the target motion model, which is used by the Gaussian
tracking algorithm in MATCH, MTST, or by its non-Gaussian tracker the
Non-Gaussian Data Fusion Module (NGDFM), to provide estimates of targets' locations
and velocities. The other models, for ELINT, classification attributes,
data rates, and false targets-play similar roles with respect
to the other target attributes.
|
MATCH can process reports in any order and is
capable of presenting information to the operator which is valid
at any selected time-before or after the "current time"
or the time of the last report. Also, the algorithm automatically
maintains multiple hypotheses in cases of ambiguous correlations,
and automatically partitions the targets into groups called "clusters,"
separated geographically or otherwise, so that different areas
of confusion can be dealt with independently. These are the fundamental
features of MATCH that set it apart from many other data fusion
algorithms. |
 |
MATCH was initially developed in Ada, has also been ported to C, and is fully tested. It is
implemented in our important data fusion systems, notably the Non-Gaussian Data
Fusion System (NGDFS), the Ground Attack Data Fusion and Optimization System
(GADFOS), the Near-Real-Time Data Fusion (NRTDF)
system, and the Global Correlation Engine.
|
