The Double Asteroid Redirection Test (DART) had an impact with Dimorphos (a satellite of the asteroid Didymos) on 26 September 2022¹. Ground-based observations showed that the Didymos system brightened by a factor of 8.3 after the impact because of ejecta, returning to the pre-impact brightness 23.7 days afterwards². Hubble Space Telescope observations made from 15 minutes after impact to 18.5 days after, with a spatial resolution of 2.1 kilometres per pixel, showed a complex evolution of the ejecta³, consistent with other asteroid impact events. The momentum enhancement factor, determined using the measured binary period change⁴, ranges between 2.2 and 4.9, depending on the assumptions about the mass and density of Dimorphos⁵. Here we report observations from the LUKE and LEIA instruments on the LICIACube cube satellite, which was deployed 15 days in advance of the impact of DART. Data were taken from 71 seconds before the impact until 320 seconds afterwards. The ejecta plume was a cone with an aperture angle of 140 ± 4 degrees. The inner region of the plume was blue, becoming redder with increasing distance from Dimorphos. The ejecta plume exhibited a complex and inhomogeneous structure, characterized by filaments, dust grains and single or clustered boulders. The ejecta velocities ranged from a few tens of metres per second to about 500 metres per second.

Accurate and robust on-board orbit determination is essential for enabling autonomous spacecraft operations, particularly in scenarios where ground control is limited or unavailable. This paper presents a novel method for achieving robust on-board orbit determination by integrating a loosely coupled GNSS/INS architecture with an on-board orbit propagator through error Kalman filtering. This method is designed to continuously estimate and propagate a spacecraft’s orbital state, leveraging real-time sensor measurements from a global navigation satellite system (GNSS) receiver and an inertial navigation system (INS). The key advantage of the proposed approach lies in its ability to maintain orbit determination integrity even during GNSS signal outages or sensor failures. During such events, the on-board orbit propagator seamlessly continues to predict the spacecraft’s trajectory using the last known state information and the error estimates from the Kalman filter, which were adapted here to handle synthetic propagated measurements. The effectiveness and robustness of the method are demonstrated through comprehensive simulation studies under various operational scenarios, including simulated GNSS signal interruptions and sensor anomalies.

In recent years, Vision-Based Navigation (VBN) techniques have emerged as a fundamental component to enable autonomous spacecraft operations, particularly in challenging environments such as planetary landings, where ground control may be limited or unavailable. Developing and testing VBN algorithms requires the availability of a large number of realistic images of the application scenario; however, these are rarely available. This paper presents a novel rendering software tool to generate accurate synthetic optical images of the lunar surface by leveraging high-resolution Digital Terrain Models (DTMs). Unlike traditional ray-tracing algorithms, the method iteratively propagates camera rays to determine their intersection with the terrain surface defined by a Digital Elevation Model (DEM). The color information is then retrieved from the corresponding Digital Orthophoto Model (DOM) through the knowledge of the ray impact points, bypassing the need for the costly computation of shadows, reflections, and refractions effects. The rendering performance is demonstrated through a comprehensive selection of images of the lunar surface under different illumination conditions and camera orientations.

To support the increasing number of planned lunar missions, a collaborative international initiative is underway to conceptualise and establish a lunar satellite constellation for communication and navigation. In this context, the goal of the current paper is to analyse what the obtainable performance is for a lunar lander that executes state estimation employing one-way ranging signals from such a Lunar Navigation Service (LNS). In particular, a small-sized optimised navigation constellation is considered as the main source of measurements, which, coupled with an accelerometer and an altimeter, is used to estimate the lander absolute trajectory during the main braking phase. The guidance is extracted on board by interpolation of a ground-optimised trajectory, followed by a reference-tracking regulator. Two alternative control tuning cases are presented, one targeting high performance, the other targeting low propellant mass. Nominal performance and associated sensitivity analyses assessed the feasibility of supporting such a critical phase with a reduced LNS constellation, reaching final control errors below 500$\mathrm{m}$, with the better performing one going down to 56$\mathrm{m}$. Among the two proposed alternatives, the one targeting low fuel expenditure has proven, however, to also be more robust against time and state uncertainty, providing much larger success rates.