A Bayesian approach to multi-target tracking and data fusion with out-of-sequence measurements

When fusing the information from several sensors, the possibility that measurements are received in the wrong order must be considered. This is especially true if the sensors are of different types, or if human observations are to be included. Recent advances in the field of Monte Carlo sampling procedures, particularly the particle filter, allow for the sequential tracking of nonlinear and non-Gaussian dynamic systems using a cloud of samples. We build on this to generate an algorithm capable of incorporating out-of-sequence measurements (OOSMs) in the general nonlinear non-Gaussian framework. The algorithm is applied to the problem of tracking where the measurements are made by a scanned sensor.