General Measurement Design Philosophy for the Kalman Filter Estimation with Special Application Given to the Imaging Radar Autofocus Update

摘要

The value of an Autofocus update of an INS is investigated. Three cases are considered: (1) centripedal acceleration only; (2) centripedal and line-of-sight acceleration; and (3) centripedal and line-of-sight acceleration as well as attitude error effects. The extended kalman filter configuration was employed using the versatile SOFE Monte Carlo simulation program. Measurement matrices were defined for each of the three cases. Simulation results indicated an observability problem for Case (1). Suggestion for further work was included. For the centripedal acceleration case only, the Autofocus measurement proved to be a useful measurement when supplemented by additional measurements. In particular, a Compass, an Autofocus, and doppler velocity update realize better than a 94% update quality. Without the Autofocus, the quality suffers greatly. (Author)