The TU Delft Astrodynamics Toolbox (Tudat) is a free open-source software (FOSS) suite geared towards research and education in computational astrodynamics. It has been applied primarily to numerical simulation of the dynamics of objects in space, ranging from optimization of re-entry vehicle dynamics to the modeling of planetary spacecraft tracking and the dynamics of natural solar system bodies. The powerful and versatile estimation module of Tudat has been used for a broad range of studies for both current and future space missions. It has the capability to combine optical and radiometric tracking data from multiple spacecraft with Earth-based observations into a comprehensive estimation of the dynamics of both natural and artificial solar system bodies, as well as physical parameters of interest. Building upon this general and adaptable framework, recent developments have focused on incorporating the necessary functionality required for real tracking data analysis. In this paper, we present the integration of these capabilities into Tudat’s fully open-source framework, with a combined focus on planetary missions and Space Situational Awareness (SSA). At present, the software provides capabilities to process several categories of observational data: (i) deep-space Doppler and range tracking data of planetary missions collected by the Deep Space Network (DSN) and ESA’s ESTRACK, supporting multiple formats such as IFMS, ODF, and TNF; (ii) deep-space Doppler and VLBI tracking data of planetary missions collected by the Planetary and Radio Interferometry and Doppler Experiment (PRIDE) with radio (astronomy) telescopes; (iii) optical astrometry and radar tracking archived by the Minor Planet Center (MPC) and the Natural Satellite Data Center (NSDC). By computing observation residuals using existing orbital solutions as references, we show that our observation models are accurate to the intrinsic quality of the data (e.g., better than 0.05 mm/s for typical deep-space Doppler data). Additionally, we demonstrate that our dynamical models possess the level of fidelity necessary to enable precise orbit estimation, effectively leveraging the high quality of the available tracking data. Tudat is unique in providing modular and flexible open-source high-fidelity modeling across a broad range of orbital regimes, enabling interdisciplinary applications. We provide an overview of the data processing and estimation capabilities and give examples from various mission domains. These include high-precision orbit estimation using deep-space Doppler tracking data, orbit determination of cis-lunar/xGEO space debris in highly non-linear regimes (specifically targeting upper stages of lunar missions) from astrometric data, and estimation of small solar system bodies using astrometric data.

During the last couple of years, there has been an increased interest to develop new deep learning networks specifically for processing 3D point cloud data. In that context, this work intends to expand the applicability of one of these networks, PointNet, from the semantic segmentation of indoor scenes, to outdoor point clouds acquired with Airborne Laser Scanning (ALS) systems. Our goal is to of assist the classification of future iterations of a national wide dataset such as the Actueel Hoogtebestand Nederland (AHN), using a classification model trained with a previous iteration. First, a simple application such as ground classification is proposed in order to prove the capabilities of the proposed deep learning architecture to perform an efficient point-wise classification with aerial point clouds. Then, two different models based on PointNet are defined to classify the most relevant elements in the case study data: Ground, vegetation and buildings. While the model for ground classification performs with a F-score metric above 96%, motivating the second part of the work, the overall accuracy of the remaining models is around 87%, showing consistency across different versions of AHN but with improvable false positive and false negative rates. Therefore, this work concludes that the proposed classification of future AHN iterations is feasible but needs more experimentation.