Abstract:

In multi-sensor data fusion system, track association is the fundamental for fusing data from different sources. However, the result of track association always contains misassociations in the presence of random errors, system biases, false tracks and track loss. Misassociations are disastrous to subsequent fusion processes. No effective misassociation detection method exists at the current state of the research area. Taking the coupling characteristics of system bias estimation and track association into consideration, the Gaussian detector and Chi-square detector for misassociations on the basis of analyzing the residual errors of system bias estimation are proposed. Simulation results show that the proposed detectors are capable of detecting misassociaitions effectively and make good prospects for real applications.