Multi-Stage Data Fusion in Security and Defence

The fundamental problem of target tracking is to estimate the state of one or more objects that persist over time, based on noisy measurements contained in a vast quantity of mostly spurious measurement data. Target tracking is closely related to a number of basic problems in statistical modelling and information extraction from noisy data. Multi-stage processing provides a wealth of processing options that can be exploited to achieve robust and high-performance surveillance. This manuscript describes a number of multi-stage tracking architectures that the author has recently studied. Additionally, we study the target cardinality problem.