Optimization-based Local Relative Navigation for Exploration on Asteroid Surface

Abstract

For navigation on asteroid surface, optimizing jointly a set of system states over a collection of measurements is an alternative to filtering-based state estimators. In order to investigate the feasibility of a sliding window smoother for the task of state estimation in exploration on the asteroid terrain surface, we propose an optimization-based relative flash LiDAR aided-inertial navigation algorithm. The local relative navigation framework is chosen for locally drift-free state estimation accounting for relative state observations from registration of flash LiDAR point clouds. State constraints derived from the point cloud registration and IMU pre-integration are leveraged in the optimization. Simulation-based validation in a high fidelity simulated environment shows that the proposed optimization-based local relative navigation is capable of estimating spacecraft system states accurately with acceptable computation complexity in our application cases.