ABSTRACT
Since there always exists some random measurement noises, there is a need to have some filtering techniques to reduce the noise effect. Kalman filter was developed in 1960’s just for this purpose. Since then, Kalman filter has been widely used in many applications, including some high profiles missions, such as Apollo project. Section 8.1 discusses the extended Kalman filter which is designed specifically for nonlinear systems including spacecraft attitude control system. Two important differences from other applications of Kalman filter in spacecraft attitude systems are introduced. First, reduced quaternion model is used to replace full quaternion model; second, additive quaternion model rather than multiplicative model is used in the extended Kalman filter. These two modifications make the filter much simpler and the computation is much more efficient. Section 8.2 provides formulas for standard Kalman filter which uses linearized spacecraft model. Although extended Kalman filters are exclusively used in spacecraft attitude determination systems because the spacecraft systems are nonlinear, the author believes that Kalman filter has its merits. Section 8.3 is a short comment on possible flight test to compare the two methods.
