Some Error Probabilities for the Association of Passive DF Measurements with Radar Returns

摘要

An important problem in multiple-sensor integration is the correct association of data from different sensors with the same target. This report considers a family of four discriminants for associating a set of measurements from a passive DF system with one of the radar targets being tracked by a surveillance radar system. This association involves choosing the radar target that will minimize a discriminant which is a function of the DF data and data on that radar target. The discriminants considered here are sums of squared differences of independent unbiased Gaussian measurements of bearings to stationary targets. Because data generally arrive from the different sensors asynchronously, it is becoming common practice with this type of system to smooth (average together) the radar data, the DF data, or both prior to use in the discriminant. This report attempts to provide a basis for deciding which smoothing option is most appropriate by giving the error probabilities for each of these three smoothing options and for the option of no smoothing.