SONAR TARGET MOTION ANALYSIS; WEIGHTED CHURN AGAINST A MANEUVERING TARGET

摘要

n a new sonar bearings-only solution method, research achieved substantial reduction in range errors on a zigging target, one-sixth those of the usual (unweighted) CHURN method. The method yields time-corrected range (value at time when expected error is least) and weights observations according to assumed zig statistics. It also advocates extrapolating favorably-chosen time-corrected ranges to obtain present range. It was found that the CHURN, with weights equivalent to those used in the research achieves equally small time-corrected range errors, and errors at solution time one-third those of the usual CHURN. Random bearing noise, however, seriously degrades solutions using these weights, even without zigs, in which case the unweighted CHURN is optimum. For combinations of zigs and bearing noise, optimum combined weighting functions exist. Unsuccessful attempts were made to use data available to the tracking ship, e. g. autocorrelation of solution residuals, for selecting optimum weighting. Autocorrelation was also probed for zig detection clues without success. Results obtained by extrapolating pairs of time-corrected range to present time were about equally as good as from single solutions using the same data. It is concluded that the important contribution of the research is to show the effectiveness of appropriate statistical weighting.