Estimation of Manoeuvring Targets using Hybrid Filters

Modern guided weapons are frequently required to operate in a complex environment that often involve highly complicated behaviours. In these types of engagements, assumptions of linearity no longer hold and model uncertainties in the form of unmeasured aerodynamic coefficients and complex non-linear aerodynamics are common. The target filter is an important sub-system of any guidance loop that extimates the required target and engagement information. To improve the performance of this target filter in challenging environments involving manoeuvring target, a fill understanding of any non-linearities present is required. The aim of this report is to investigate the manoeuvring target filtering problem, to examine the importance of mode measurements and to examine a particular filtering approach. A review of existing filtering results is provided before introducing a non-linear filtering approach known as hybrid filtering. Three possible filtering approaches are examined in simulation studies: the extended Kalman filter, the interacting multiple model filter and a new hybrid filtering approach. The simulation studies suggest that mode measurements may improve target filtering performance but the studies do not support the use of the examined hybrid filtering approach. Some refinement of this hybrid filtering approach is required. An improved understanding of filtering techniques is required to aid support of present upgrade programs involving the guidance loops of new air-to-air and standoff missile systems. This understanding is necessary for the support of future weapon procurement upgrade programs