| 摘要 |
The invention relates to a method for correlating target tracks (14) from multiple track sources into a single situation picture (15). Here the track sources are comprised of various sensors (11) and target tracking systems (12). The track correlator (13) embodies the new correlation method, which, unlike previous art, is based on correlating the tracks (14) to target numbers. A large number of correlation hypotheses are evaluated along with the probability for each hypothesis. The method runs recursively: existing probabilities are propagated until the current point of time, and then new probabilities are computed, e.g. using Bayes' rule, according to how well each hypothesis fits with the observed track data. <??>To reduce the number of hypotheses that have to be handled individually, a simplified scheme is introduced where each track (14) is characterized by one (or more) target number and the corresponding probability, and where each target number is characterized by the probability that it represents a real and trackable target. The simplified representation is particularly useful for distributed correlation, where a number of correlators run at different sites, to form a common situation picture with common track numbers. Further, as the number of hypotheses can be very large, methods for just evaluating the most plausible hypotheses have been suggested. <IMAGE>
|
