Observer based input estimation approaches for seeker filtering

The paper investigates the possibilities of filtering base motion disturbance in a missile seeker from the target sight line rate (guidance) signal using observer based approaches. Unlike conventional Kalman filter based approaches, where the target kinematics is assumed as the plant description, the seeker dynamics with its track and stabilization loops are considered as the plant and the platform motion is considered as a known input. Two observer-based approaches are described. In order to make the complex plant with nonlinearities and delay amenable to the linear observer theory, a simplified essential model of the seeker is used. First, the concept of a high-gain observer (JHG) and its use as an estimator of unknown input are introduced. It has been shown that with proper choice of observer excitation signals and a robust pole assignment technique, a very simple observer structure is possible - which neither needs the platform motion signal nor is sensitive to stabilization loop parameters and nonlinearities. Further, it is also possible to reduce the effective lag due to the track loop. A PI-type state observer, called JPI, is studied next where the target sight line rate (SLR) becomes a state, rather than the input. A simple trade-off mechanism is introduced which balances the requirements for fast tracking and sensitivity to measurement noise, in a manner akin to Kalman filtering. Finally, the relative merits of the modified observer based seeker filters as compared to conventional loop decoupling (Garnell Type) approaches are discussed. It is argued that the type of filtering proposed may relax the electromechanical base motion isolation requirement.