Real-time state estimation for asteroid close-proximity operations via lidar altimetry and a particle filter

Current practice for asteroid close proximity operations typically initiates a maneuver from a safe orbit, and then the maneuver is executed open loop, i.e., once the maneuver is initiated, an estimate of the spacecraft's state is not used to correct for errors. A problem with this approach is that extremely accurate modeling of the asteroid's dynamics is required. Moreover, even with accurate modeling, there is still often considerable error between the targeted and actual spacecraft state at the end of the maneuver, making precision touch and go (TAG) and landing maneuvers impossible. Without real-time spacecraft state estimation that can be coupled to a guidance law, there is no way to improve on this state of affairs. This paper demonstrates how such a real-time state estimation algorithm can be constructed using a Rao-Blackwellized particle filter, a laser altimeter, and an asteroid shape model. The state estimation algorithm is coupled with a guidance law, and precision TAG maneuvers on Itokawa and RQ36 are demonstrated via Monte Carlo simulations in a high fidelity simulation environment.