Traditional space cataloguing approaches first combine consecutive measurements into "tracklets" and then associate these tracklets definitively to known objects, before updating the orbits accordingly. In contrast, multi-object tracking (MOT) methods consider multiple measurement association hypotheses simultaneously, but typically discard the pre-formed tracklets and avoid conclusive measurement-to-object assignments. This thesis presents a flexible MOT library based on finite set statistics (FISST) and proposes a modified multi-object filter that fully leverages the FISST framework while incorporating prior knowledge from existing tracklets. Additionally, a robust method is introduced to extract the most probable measurement assignments directly from the filtering recursion, enabling targeted single-object post-processing. The new tracklet filter is demonstrated to effectively discover and maintain state estimates for objects in low Earth orbit and geosynchronous orbit, using sparse optical observations from both ground-based and space-based telescopes with diverse pointing strategies.

Nowadays, space-based systems for navigation, communication, Earth observation, meteorology and many other applications are indispensable for services critical for society. In order to preserve these vital applications among the increased risks in space, the Space Surveillance and Tracking segment of the Space Situational Awareness program has been set up in the last decade by the European Space Agency. This segment is occupied with one core activity: maintaining and building up a space object catalog. The object states within it can be used for conjunction assessment, collision avoidance and fragmentation event detection, allowing active satellites to be protected.

For objects in Low Earth Orbit, the Space Surveillance and Tracking segment primarily employs radar, which uses active signals, whose signal strength degrades with the fourth power of the distance. Beyond this orbit regime, passive optical sensors are primarily used, because they do not require an active signal and are therefore able to observe beyond Low Earth Orbit without large power requirements. As a consequence of the sparse optical sensors which have to track the large population of space objects, the optical observing arcs are typically very short: minutes at most. These short optical arcs are called tracklets.

Determining an orbit from a short arc is a challenging problem and is commonly referred to as the too short arc problem. Specifically, this problem entails that a single short arc tracklet is insufficient to construct an orbit with realistic state estimates. In order to obtain accurate state estimates, the object corresponding to the tracklet must be re-observed during a later pass to obtain six independent parameters of high quality. The initial problem to solve, however, is to identify which tracklets are correlated and thus associated with the same object. Thus, tracklet correlation is a critical step in the cataloging of objects. Evaluating and improving the performance of tracklet correlation methods is the key subject of this thesis. The performance is assessed on three key metrics: the association accuracy, the computational efficiency and the robustness.

In the first part of this thesis, tracklet correlation methods are evaluated. The most promising ones have been identified to be the Adapted Gooding method, the BVP optimization and IVP optimization. The performance of these methods is evaluated for different data distribution characteristics and orbit regimes, first without incorporating perturbations in the tracklet correlation methods.

In the second part, the tracklet correlation methods are modified to improve the performance. A major gap in the state of the art has been successfully addressed, namely to incorporate perturbations in the tracklet correlation methods. Results have shown a significant increase in association accuracy. Specifically, tracklet pairs with arbitrary time gap can be associated with the same true positive rate. Moreover, tracklet pairs from different orbit regimes can be associated with a similar true positive rate if the most important perturbations are incorporated.

Finally, the BVP and IVP optimization are found to be robust to noise, while the Adapted Gooding method severely suffers from the perfect measurement assumption. The performance of the latter has been improved by improving the accuracy of the angles through a Least Squares fit. Although the BVP and IVP optimization are robust to noise, the performance is limited by the accuracy of the angle rates, according to the Chi-squared distribution. A novel tracklet correlation method called the BVP-DC2-DC3 is developed to overcome this limitation, achieving an excellent TPR of 100% and a TNR of 99% for the analyzed objects in GEO for tracklets with a measurement noise of one arcsecond. This accuracy is superior to literature. Moreover, the computational efficiency is superior or comparable to literature. Additionally, the BVP-DC2-DC3 has a high robustness, because for the analyzed GEO tracklets with a measurement noise of five arcseconds, a TPR of 100% and a TNR of 95% can still be achieved.

JUICE will measure tidally-induced variations in Ganymede's gravitational field with unprecedented accuracy. Previous research focused on studying tides with a spherically-symmetric interior allowing to constrain the mean shell thickness and subsurface ocean. The presence of lateral variations in the interior alters the tidal response and produces additional signals. This research investigates whether JUICE measurements will be precise enough to constrain possible lateral variations in Ganymede. The structure of Ganymede's ice shell is crucial to the question of habitability, since it influences the process of tidal heating and the transport of heat and material in the interior. This research uses simulated range and Doppler observations during JUICE's orbital phase, to obtain the expected uncertainties in measuring the tides and compares them to the expected signals for a laterally heterogeneous Ganymede model. Results indicate that JUICE could detect zonal lateral variations of degree 1 and 2 if their amplitude is at least ~5%

The traditional approach to space mission design considers deterministic dynamics, relying on a Guidance, Navigation and Control (GNC) system to manage uncertainty dispersions. To simplify the iterative process of trajectory design, reduce the burden on GNC, and increase mission safety, this thesis investigates the use of uncertainty propagation to robustify a deterministically-optimal atmospheric entry trajectory. A hybrid Gaussian Mixture Model-Polynomial Chaos Expansions method is optimised and used to generate analytical expressions of the entry corridor statistics as a function of 16 uncertainty sources and five decision variables. These expressions reduce the design space and directly handle stochastic constraints, simplifying the subsequent optimisation. The method is validated on the HORUS-2B vehicle and reference trajectory, widening the entry corridor margins to 2-sigma bounds with errors in the order of 2%. The output of this thesis is a generic stochastic optimisation methodology that is more efficient than traditional optimisation with Monte Carlo simulations.

This thesis presents a methodology to characterize the interior structure of Ganymede, Jupiter’s largest moon, through a joint Bayesian inversion framework that integrates gravity, magnetic induction, tidal, and libration observations. The motivation arises from the growing scientific interest in icy moons, which not only provide insights into Solar System evolution but also represent promising candidates for habitability due to their subsurface oceans. Ganymede stands out as the largest moon in the Solar System and the only one known to possess its own intrinsic magnetic field. It is also believed to host a subsurface ocean beneath its icy crust, making it an interesting target for scientific exploration. However, much of its internal structure remains uncertain. The upcoming ESA’s Juice mission will carry out detailed observations of the moon through a series of flybys followed by an extended orbital tour, delivering high-precision data that will help constrain Ganymede's structure.
This work addresses the challenge of constraining Ganymede's interior structure using multiple datasets - gravity, magnetic induction, tidal, and libration observations - each sensitive to different interior parameters and affected by parameter degeneracies. We first perform a global sensitivity analysis to identify how each observable relates to specific interior properties. Our model assumes a five-layer spherical structure comprising a metallic core, silicate mantle, high-pressure ice, liquid salty ocean, and outer ice shell. The sensitivity analysis shows that magnetic induction is most sensitive to ocean thickness and composition, tidal displacement to ice shell thickness and rigidity, and libration amplitude to shell rigidity. Degeneracies also emerge, such as between shell thickness, ocean density, and shear modulus, highlighting the need for a joint inversion approach.
The Bayesian inversion is carried out progressively. Starting with Ganymede’s moment of inertia, we constrain the core and mantle but leave the hydrosphere largely unconstrained. Including magnetic induction data substantially improves the estimation of ice shell and ocean thicknesses, while the real part of the tidal Love number k₂ refines constraints on the densities of the ice shell and ocean and enables inference of ocean composition. These observations also provide information on the rigidity of high-pressure ice and the compressibility of interior layers. Finally, the combination of moment of inertia, magnetic induction amplitude, and both the real and imaginary parts of k₂ offers constraints on the viscosities of the ice layers, related to the tidal dissipation within Ganymede.
Beyond advancing our understanding of Ganymede, this study demonstrates the effectiveness of joint Bayesian inference for the characterization of planetary interior.This framework contributes to the scientific preparation for the ESA's Juice mission, identifies the most critical measurements to break degeneracies in interior model parameters, and can be extended to other icy moons, such as Europa and Enceladus, where spacecraft data will provide similar observational constraints.

Due to the increasing interest of the aerospace industry and the scientific community in missions targeting halo orbits, a specific family of periodic solutions within the circular restricted three-body problem, it is highly desirable to find ways to reduce the cost of transfers between these orbits. To achieve this, depending on the mission characteristics, one could opt for reducing the required propellant mass or the time of flight. As such, with two objectives to be minimized, an ideal implementation for a mission designer would provide a Pareto front of feasible transfers instead of the single trajectory commonly obtained with conventional methods. Moreover, the necessary propellant mass can be minimized even further by means of electric low-thrust propulsion due to the significantly larger exhaust velocities. Therefore, the purpose of this research is to develop, implement, and study a suitable approach to obtain a collection of low-thrust transfers between halo orbits optimized in terms of both propellant mass and time of flight.

The approach employs an optimal control indirect method as thrust law which, combined with a heuristic optimizer based on differential evolution, can find a wide variety of trajectories that minimize the aforementioned objectives within the circular restricted three-body problem. Heuristic optimization is employed to remove the dependency of the solution on the provided initial guess and find trajectories in the region of the global minima. To satisfy the demanding boundary conditions characteristic of indirect methods, these constraints are included as a third objective for the optimizer to minimize as well. Due to the tolerance allowed on the constraints, the trajectories are subsequently refined with direct collocation methods. Then, they can be transitioned to a high-fidelity model, taking advantage of the versatility of direct methods. Furthermore, the implementation allows for the inclusion of invariant manifold phases arising from the departure and target orbits to obtain a wider set of Pareto-optimal solutions.

To assess the suitability of the proposed procedure, a specific transfer between two halo orbits around different Lagrange points of the Earth-Moon system was optimized. The results consisted of 100 Pareto-optimal transfers spanning more than 60 days, offering significant mission design freedom. Moreover, the Pareto front outperforms the trajectory found in the literature for a comparable use case by 30\% in all objectives. Next, an optimized trajectory was first successfully verified with the mission analysis software ASTOS and then refined with direct collocation. The refined trajectory exhibited negligible changes in performance, rendering the trajectories obtained with this approach as promising initial guesses for further optimization with direct collocation methods. Moreover, several major perturbations not included in the simplified dynamic system were also corrected with direct collocation. The next steps include: accounting for the eccentricity of the Moon's orbit to fully transition the trajectories to a high-fidelity model, assessing the optimization quality with different use cases, and implementing transfers between different periodic solutions and dynamic systems, such as the Sun-Earth system.

Cislunar space is getting increasingly important. Countries like the US and China are directing their attention towards the Moon with the Artemis and Chang’e missions. Simultaneously, space debris pollution of our orbits is escalating, increasing the risk of the Kessler syndrome occurring. Space Situational Awareness (SSA) aims to prevent this. The sheer amount of space debris makes continuous tracking unfeasible. This creates the need for accurate orbit determination and propagation between observations. Model frameworks have been developed extensively for space debris in near-Earth orbits, but there is little experience with cislunar space. This region is more challenging because of its unstable nature, increasing the risk of losing track of objects.

In this research, a model framework is developed, using the open-source Python package Tudat, that can estimate and propagate long-term cislunar space debris orbits accurately from optical data, whilst quantifying the uncertainties realistically over time using Monte Carlo simulation. Orbit determination has been performed using Weighted Least-Squares. The algorithm can estimate (amongst other parameters) initial states of the objects, model parameters like radiation pressure properties, and observation bias. We apply our framework to the Chang’e 2 and 3 upper stages. A selection of 13 estimation windows has allowed for analysis on a diverse range of orbital and observation characteristics. Moreover, the effect of close Moon approaches on orbit determination quality and propagation accuracy has been investigated.

A generic model framework has been found that achieves sufficient accuracy with reasonable computational load for all 13 use cases and can be used as a foundation for other cislunar space debris studies. The generic framework only estimates initial state, and uses a relatively simple dynamical model and integrator configuration. This allows sufficiently accurate propagation up to 2 years out-of-sample, depending strongly on the stability of the orbit. Afterwards, the generic model framework is tailored on individual use cases to improve its performance. Parameters like the radiation pressure coefficient and observation bias are estimated in this process. The tailored model framework improved performance for 8 out of 13 use cases (compared to the generic model framework), decreasing out-of-sample RMSE between 20-95% and increasing period of sufficient accuracy with up to 250 days. Estimating on 7-10 months of observations results in the best orbit determination quality and propagation accuracy for the cislunar use cases. The tailored model framework performs robustly for various non-linear orbits, except for out-of-sample close Moon approaches. Several solutions are proposed to solve this issue. Finally, it is found that the effect of uncertainty for cislunar space debris orbits over time is significant in the current framework. Uncertainty over time is especially large when estimating on short estimation windows (<4 months) and for orbits experiencing non-linear behaviour.

The congestion of orbits around the Earth, particularly in Low Earth Orbit (LEO), necessitates the implementation of space-based Space Situational Awareness (SSA) missions. This increasing orbital congestion, driven by the proliferation of mega constellations, poses significant challenges to satellite operations and collision avoidance. In response, commercial entities are beginning to plan and launch space-based SSA missions. Despite this emerging interest, there is a notable gap in the literature regarding the scheduling of operations for these missions, which are characterised by the inclusion of multiple operational modes.

Additionally, existing literature on operational scheduling often lacks high-fidelity modelling of onboard resource dynamics. This research addresses this gap by focusing on energy-constrained scenarios, developing novel operational scheduling frameworks suitable for assigning data collection and data downlink tasks for optical space-based SSA missions under such constraints.

To enhance the fidelity of the scheduling process, this research employs both low-fidelity and high-fidelity simulators to model the dynamic behaviour of onboard resources. The developed frameworks are tested in various scenarios to validate their practicality and effectiveness. A significant contribution of this research is the open-source availability of part of the developed scheduling frameworks, utilising software accessible under the TU Delft license. This open-source approach ensures that the research can be extended and built upon by future researchers, fostering ongoing advancements in the field of space-based SSA operations.

Performing crucial activities for space exploration, e.g., debris removal and on-orbit servicing, systems for Rendezvous and Proximity Operations (RPO) are required to be autonomous and scalable. Within this context, learning-based relative navigation has gained significant traction due to the latest advancements in AI and the cost-effectiveness of monocular cameras.

This thesis introduces a pose estimation system composed of an Object Detection (OD) and a Keypoint Detection (KD) network for keypoint extraction, coupled with a pose solver, and trained purely on synthetic images. When applying realistic data augmentations, the system achieves a reduction in KD error by 80%, further improved by training on photorealistic images. After an architecture optimization step, the final system consistently meets the inference time requirements on the Myriad X edge processor. This proves the feasibility of developing RPO systems with artificial data, showcasing a scalable approach, while complying with the limitations of onboard hardware.

The Mars Ascent Vehicle will bring samples back from the surface of Mars to its low orbit in 2031. Its preliminary design has been performed by NASA, defining it as a two-stage solid propellant rocket. However, a research gap exists in finding what the optimum solid rocket motors geometries are to bring the Martian samples to a target orbit between 300km and 375km above Mars.
To this extent, the ascent and propellant burn have been modelled, allowing for variation in the geometry of the motors, but also in the launch and separation angles, and in thrust vectoring control. This model has been incorporated into an optimisation, from which a minimum lift-off mass of 343kg was found, 57kg less than required. In the process, 18 thousand different solutions have been found that satisfy all requirements, would the priority shift from the mass of the vehicle to for instance the burn time, or the final altitude.

Single-Stage-To-Orbit (SSTO) launch vehicles are capable of reaching orbit while being potentially fully reusable. A technology capable of enabling SSTO access to space is the air-breathing engine, which has a high specific impulse and low thrust-over-weight (T/W) ratio compared to conventional rocket engines. This necessitates the use of a Horizontal Take-off and Horizontal Landing (HTHL) launch vehicle with a lifting surface, which enables banking maneuvers that can manipulate the orbital Right Ascension of the Ascending Node (RAAN) that the space plane achieves. This means that the space plane is capable of advancing or delaying its time of launch, while still achieving a similar orbital RAAN. This effectively increases the launch window even if a precise orbit insertion is necessary.

This thesis answers the question what the fuel-optimal ascent trajectory is for a space plane that would extend its launch window. In order to answer this question, the full six DoF equations of motion have been defined to simulate the ascent trajectory of a space plane. Additionally, a robust Guidance and Control system is developed that can use the nonlinear translational and rotational equations of motion.

For several launch times, ranging from two hours earlier to two hours later than the nominal launch time, the ascent trajectory has been optimized that included a banking maneuver. It was found that at the cost of propellant, it is possible to have a launch time delayed are advanced by as much as two hours, which means that the space plane can manipulate the RAAN by ± 30 degrees. The extra propellant cost can be related to increased drag losses during the banking maneuver.

Interplanetary missions have relied on Radio Science (RS) for recovering gravity fields via detecting their signature on the spacecraft’s trajectory. Yet the weak gravitational fields of small bodies, coupled with the prominent influence of confounding accelerations, hinder the efficacy of this method. Meanwhile, quantum sensors based on Cold Atom Interferometry (CAI) have demonstrated absolute measurements with an inherent stability and repeatability, reaching the utmost accuracy in microgravity. This work addresses the potential of CAI-based Gradiometry (CG) as a means to strengthen the RS gravity experiment for small-body missions. Phobos represents an ideal science case as astronomic observations and recent flybys have conferred enough information to define a robust orbiting strategy, whilst promoting studies linking its geodesic observables to its origin. A covariance analysis was adopted to evaluate the contribution of RS and CG in the gravity field solution, for a mission duration of one week.

The favourable observational geometry and the small characteristic period of the gravity signal add to the competitiveness of Doppler observables. Provided that empirical accelerations can be modelled below the nm/s2 level, RS is able to infer the 6x6 spherical harmonic spectrum to an accuracy of 0.1-1% w.r.t. the homogeneous interior values. If this correlates to a density anomaly beneath the Stickney crater, RS would suce to constrain Phobos’ origin. Yet, in event of a rubble pile or icy moon interior (or a combination thereof) CG remains imperative, enabling an accuracy below 0.1% for most of the 10x10 spectrum. Nevertheless, technological advancements will be needed to alleviate the current logistical challenges associated with CG operation. This work also reflects on the sensitivity of the candidate orbits with regards to dynamical model uncertainties, which are common in small-body environments. This brings confidence in the applicability of the identified geodetic estimation strategy for missions targeting other moons, particularly those of the Giant planets, which are targets for robotic exploration in the coming decades.

What is the most advantageous magnetic attitude control torquer for reducing angular rates of a PocketQube? To answer this question several magnetorquer concepts are explored and two concepts are built. In addition, a method to test these magnetorquers is developed. An Air-core magnetorquer system and a magnetorquer embedded in a PCB are designed, built and tested. By building and testing the designs also possible construction issues related to a concept can be found, which are not evident when only designing the magnetorquers. The Air-core magnetorquers also serves as a prototype for the magnetorquer for Delfi-PQ, the satellite to be launched by Delft University of Technology. After testing two concepts are then compared with each other and also with other concepts such as a ferromagnetic torque rod.

The control of aircraft can be carried out by Reinforcement Learning agents; however, the difficulty of obtaining sufficient training samples often makes this approach infeasible. Demonstrations can be used to facilitate the learning process, yet algorithms such as Apprenticeship Learning generally fail to produce a policy that outperforms the demonstrator, and thus cannot efficiently generate policies. In this paper, a model-free learning algorithm with Reinforcement Learning in the loop, based on Apprenticeship Learning, is therefore proposed. This algorithm uses external measurement to improve on the initial demonstration, finally producing a policy that surpasses the demonstration. Efficiency is further improved by utilising the policies produced during the learning process. The empirical results for simulated quadrotor control show that the proposed algorithm is effective and can even learn good policies from a bad demonstration.

The use of Taylor Series Integration (TSI) for the propagation of spacecraft trajectories and its combination with the relatively unknown Unified State Model (USM) coordinate system, resulting in the TSI-USM propagator, is known to allow fast propagation of interplanetary low-thrust trajectories. Unfortunately, the TSI-USM propagator is a problem-specific method, which is detrimental for its use in evolutionary algorithms.

In this thesis, a method was developed to generalize the TSI-USM propagator. The TSI-USM was generalized to allow the propagation of a spacecraft trajectory in a central gravitational field subject to any pre-set thrust profile, without the need to adapt the internal working of the propagator. The three-dimensional thrust profile should be specified a priori, and with respect to time, by the user or by an evolutionary algorithm. The method used to generalize the TSI-USM requires the coefficients of the cubic spline interpolation of the pre-set thrust profile to compute the recurrence relations of the TSI-USM.

Results show that, although the method indeed generalizes the TSI-USM, it is less accurate and requires more CPU time for a given accuracy than existing RK8(7)13M-based propagators. From this result, it can be expected that future attempts to generalize the TSI-USM propagator, without sacrificing its accuracy and computational speed, will be challenging.