Horizon-based autonomous navigation and mapping for small body missions



Journal Title

Journal ISSN

Volume Title



This report expands upon a previously developed approach to simultaneously estimate asteroid physical characteristics and relative spacecraft state with limited prior knowledge using optical observations of the illuminated horizon from resolved imagery. The approach is intended for eventual autonomous use onboard a spacecraft. The asteroid surface is represented as a star-convex shape where the radial extent is a function of the input spherical coordinates. This unknown radial extent function is modeled as a Gaussian Process, which is formulated as a state space model that is well-suited to sequential Bayesian inference methods, namely, the Extended Kalman Filter. Early versions of this algorithm solidly demonstrated proof-of-concept, but this works aims to adjust and refine the filter equations to increase robustness to larger nonlinearities. Efficacy of the reworked estimator is demonstrated through Monte Carlo simulations.



LCSH Subject Headings