A Comparison of the Boostrap Particle Filter and the Extended Kalman Filter and the Effects of Non-linear Measurements in Spacecraft Attitude Determination Using Light Curves

Orbital ATK Conference Center

The attitude of a space object can be obtained from looking at the photometric light curves of the object. This approach has traditionally been employed by astronomers for asteroids, and has recently been developed for application to satellites. Estimation algorithms such as particle filters or Kalman filters are employed to acquire the attitude information from the light curves through a process called light curve inversion. The use of a bootstrap particle filter and an extended Kalman filter in light curve inversion are examined. In many instances the extended Kalman filter is able to match the performance of the bootstrap particle filter. However, there are some cases where the extended Kalman filter diverges. The reasons leading to this filter divergence are investigated.