Abstract: Based on an analysis of the characteristics of space optical surveillance, this article proposes a modified sphere coordinatesSigma points Kalman filter (MSC-SPKF) approach to realize the threedimensional tracking of a space object which makes use of the state transformation of Sigma points. This approach utilizes the advantages of the modified spherical coordinates on bearing-only tracking and the unscented transformation (UT) of secondorder approximation of nonlinear systems. The Sigma points are selected in the modified spherical coordinates, and then transformed to Cartesian coordinates to implement the nonlinear prediction. Finally, filter updating is accomplished in the modified spherical coordinates. The corresponding tracking initiation method is also provided. Simulation results indicate that this approach has the advantage of rapidity of convergence and stability over existing methods.

Key words: space object, spacebased optical surveillance, unscented transformation, Kalman filter, tracking