1,721,036 research outputs found
Design and verification of Guidance, Navigation and Control systems for space applications
Full text linkIn the last decades, systems have strongly increased their complexity in terms of number of functions that can be performed and quantity of relationships between functions and hardware as well as interactions of elements and disciplines concurring to the definition of the system. The growing complexity remarks the importance of defining methods and tools that improve the design, verification and validation of the system process: effectiveness and costs reduction without loss of confidence in the final product are the objectives that have to be pursued.
Within the System Engineering context, the modern Model and Simulation based approach seems to be a promising strategy to meet the goals, because it reduces the wasted resources with respect to the traditional methods, saving money and tedious works. Model Based System Engineering (MBSE) starts from the idea that it is possible at any moment to verify, through simulation sessions and according to the phase of the life cycle, the feasibility, the capabilities and the performances of the system. Simulation is used during the engineering process and can be classified from fully numerical (i.e. all the equipment and conditions are reproduced as virtual model) to fully integrated hardware simulation (where the system is represented by real hardware and software modules in their operational environment). Within this range of simulations, a few important stages can be defined: algorithm in the loop (AIL), software in the loop (SIL), controller in the loop (CIL), hardware in the loop (HIL), and hybrid configurations among those.
The research activity, in which this thesis is inserted, aims at defining and validating an iterative methodology (based on Model and Simulation approach) in support of engineering teams and devoted to improve the effectiveness of the design and verification of a space system with particular interest in Guidance Navigation and Control (GNC) subsystem. The choice of focusing on GNC derives from the common interest and background of the groups involved in this research program (ASSET at Politecnico di Torino and AvioSpace, an EADS company). Moreover, GNC system is sufficiently complex (demanding both specialist knowledge and system engineer skills) and vital for whatever spacecraft and, last but not least the verification of its behavior is difficult on ground because strong limitations on dynamics and environment reproduction arise.
Considering that the verification should be performed along the entire product life cycle, a tool and a facility, a simulator, independent from the complexity level of the test and the stage of the project, is needed. This thesis deals with the design of the simulator, called StarSim, which is the real heart of the proposed methodology. It has been entirely designed and developed from the requirements definition to the software implementation and hardware construction, up to the assembly, integration and verification of the first simulator release. In addition, the development of this technology met the modern standards on software development and project management. StarSim is a unique and self-contained platform: this feature allows to mitigate the risk of incompatibility, misunderstandings and loss of information that may arise using different software, simulation tools and facilities along the various phases. Modularity, flexibility, speed, connectivity, real time operation, fidelity with real world, ease of data management, effectiveness and congruence of the outputs with respect to the inputs are the sought-after features in the StarSim design. For every iteration of the methodology, StarSim guarantees the possibility to verify the behavior of the system under test thanks to the permanent availability of virtual models, that substitute all those elements not yet available and all the non-reproducible dynamics and environmental conditions. StarSim provides a furnished and user friendly database of models and interfaces that cover different levels of detail and fidelity, and supports the updating of the database allowing the user to create custom models (following few, simple rules). Progressively, pieces of the on board software and hardware can be introduced without stopping the process of design and verification, avoiding delays and loss of resources.
StarSim has been used for the first time with the CubeSats belonging to the e-st@r program. It is an educational project carried out by students and researchers of the “CubeSat Team Polito” in which StarSim has been mainly used for the payload development, an Active Attitude Determination and Control System, but StarSim’s capabilities have also been updated to evaluate functionalities, operations and performances of the entire satellite. AIL, SIL, CIL, HIL simulations have been performed along all the phases of the project, successfully verifying a great number of functional and operational requirements. In particular, attitude determination algorithms, control laws, modes of operation have been selected and verified; software has been developed step by step and the bugs-free executable files have been loaded on the micro-controller. All the interfaces and protocols as well as data and commands handling have been verified. Actuators, logic and electrical circuits have been designed, built and tested and sensors calibration has been performed. Problems such as real time and synchronization have been solved and a complete hardware in the loop simulation test campaign both for A-ADCS standalone and for the entire satellite has been performed, verifying the satisfaction of a great number of CubeSat functional and operational requirements.
The case study represents the first validation of the methodology with the first release of StarSim. It has been proven that the methodology is effective in demonstrating that improving the design and verification activities is a key point to increase the confidence level in the success of a space mission
In Orbit Operations of an Educational Cubesat: the e-st@r-II Experience
No full textCubeSats have achieved a place of relevance in the modern space missions. However, CubeSats need improvements in terms of new technologies, mission and system reliability, and management of the entire product life cycle. A major lack of knowledge exists about the operations of past and current CubeSats in orbit. Operations planning and re-planning, failures and anomalies management, step-by-step procedures and rules of execution are crucial but poorly analyzed aspects that can increase the success rate of a mission. Moreover, information about the programmatic and technical achievements and lessons learned during the operation phases is often not available or incomplete. However, the sharing of the results, especially for educational programs, becomes fundamental to improve the quality of missions and to prevent mistakes. This paper aims at reducing this gap of knowledge and at sharing the experience and the lessons learned gained by the CubeSat Team of Politecnico di Torino over the three years of in-orbit operations of the E-ST@R-II mission. The flexible approach adopted for the operations planning, the management of anomalies and the re-planning of the operations are presented. Simple yet effective tools have been developed for mission planning and root causes identification, and they are presented through the discussion of the main technical and educational results. The paper provides also a list of good practices and recommendations applicable to future CubeSat missions
Deep Learning-Optimized Monocular Navigation for Autonomous Rendezvous and Proximity Maneuvers in Small Satellite Missions
Full text linkAccurate estimation of the position and orientation
of a spacecraft during proximity operations—such as rendezvous,
docking, on-orbit servicing (OOS), and active debris removal
(ADR)—is critical to ensuring mission success and safety. Tradi-
tional visual navigation methods based on hand-engineered fea-
ture matching often struggle with robustness and generalization,
while existing deep learning approaches face limitations due to
heuristic hyperparameter tuning and limited training data. In
this work, a novel convolutional neural network (CNN)-based
architecture for monocular pose estimation of non-cooperative
spacecraft is proposed, specifically designed to improve robust-
ness across diverse operational scenarios. The model is trained on
a high-fidelity synthetic dataset comprising approximately 25,000
images, simulating realistic proximity conditions with variations
in lighting, background textures, and spacecraft geometries.
To assess its performance, an extensive benchmarking study is
conducted against representative State-of-the-Art methods using
standardized evaluation metrics and controlled test conditions.
The results demonstrate the competitive performance of the
proposed method and provide critical insights into the factors
affecting pose estimation accuracy in realistic spaceborne appli-
cation
Hardware in the loop test campaign for e-st@r cubesat
No full textThe paper describes the Hardware-In-the-Loop (HIL) simulation methodology used in the development of the e-st@r cubesat, which is one of the cubesats chosen by the ESA Education Office for the Vega Maiden Flight. The e-st@r program is carried out by students and researchers of the Department of Mechanical and Aerospace Engineering at the Politecnico di Torino, and the cubesat has been successfully launched into orbit in February 2012.
The HIL methodology has been applied to the space segment composed by a payload (an Active Attitude Determination and Control System) and a satellite bus (an On Board Computer, an Electrical Power System and a Communication System). The simulation tests campaign objective is to investigate and evaluate the e-st@r performances during its operative life by including the main hardware of the satellite in the loop. Simple and very low cost solutions must be taken into account in order to satisfy the requirements and the constraints of the e-st@r mission. HIL simulation includes the models of 1) sensors (an Inertial Measurement Unit and a Magnetometer), 2) actuators (three magnetic torquers), 3) solar panels (five couples of GaAs TJ solar cells), 4) thermal behaviour of the satellite, 5) orbit, 6) dynamics and kinematics of the satellite.
The satellite behaviour during all mission phases, the uplink and downlink communications, and the performances (such as pointing accuracy and orbital manoeuvre capability, batteries charge/discharge time, power consumption) are the main features and high level functional requirements investigated and tested during the HIL verification campaign.
Data acquired during the tests both by the simulator PC and by the Ground Station allow to compare the hardware behaviour and the simulated response obtained from the global simulation model. The results of the verification by means of the HIL strategy are consistent with the expected values in any operative condition, thus validating the methodology. Moreover, it has been verified that testing via HIL simulations may efficiently support the design and verification of a small satellite program, reducing the time and the cost of the development phase, while at the same time increasing the effectiveness and reliability of the satellite.
The methodology tested on the e-st@r cubesat may be tailored also to other similar projects, thanks to its versatility given by its inherent modular structure
Investigation of a CubeSat in Orbit Anomaly through Verification on Ground
Full text linkGiven the role of Cubesats in the new space economy, a statistically relevant number of CubeSats have flown, and considering the high percentage of failed missions, the investigation of in-orbit anomalies becomes of paramount importance. It is rare to find data about mission failures, probably because the partial or total absence of telemetry does not encourage any analysis. The lack of data from the spacecraft in orbit can be mitigated through ad-hoc verification campaigns on satellite models when in-orbit anomalies are experienced. This paper shows an effective testing activity conducted on models of the spacecraft to understand the root cause of a severe anomaly that occurred during mission operations. The tests are part of a comprehensive methodology for root causes analysis. The paper aims at sharing the experience built upon a practical case of interest. More importantly, this work has the ambition of fostering the research on key topics of reliability, mission operations and assembly, and integration and verification/test processes, which have shown to be critical. The activity presented in this paper demonstrates that investigating the anomalies can help recover the mission of interest but can also support building a heritage that is still missing for CubeSat missions today
Verification of a CubeSat via Hardware-in-the-loop Simulation
Full text linkThis paper describes the Hardware-In-the-Loop (HIL) simulation methodology used for the verification of functional requirements of e-st@r-I CubeSat. The satellite's behavior has been investigated via HIL simulation, and the results obtained are consistent with the expected values in any operative conditions. It is proven that HIL simulation is a valuable means for supporting the verification process of small satellites and may help reduce time and cost of the development phase and increase mission reliabilit
Orthogonal-Array based Design Methodology for Complex, Coupled Space Systems
Full text linkThe process of designing a complex system, formed by many elements and sub-elements interacting between each other, is usually completed at a system level and in the preliminary phases in two major steps: design-space exploration and optimization. In a classical approach, especially in a company environment, the two steps are usually performed together, by experts of the field inferring on major phenomena, making assumptions and doing some trial-and-error runs on the available mathematical models. To support designers and decision makers during the design phases of this kind of complex systems, and to enable early discovery of emergent behaviours arising from interactions between the various elements being designed, the authors implemented a parametric methodology for the design-space exploration and optimization. The parametric technique is based on the utilization of a particular type of matrix design of experiments, the orthogonal arrays. Through successive design iterations with orthogonal arrays, the optimal solution is reached with a reduced effort if compared to more computationally-intense techniques, providing sensitivity and robustness information. The paper describes the design methodology in detail providing an application example that is the design of a human mission to support a lunar base
Algoritmi per il controllo di flotte di piccole piattaforme aerospaziali e verifica sperimentale
No full textCNN-Based Visual Navigation: Optimization Strategies for Monocular Pose Estimation in Proximity Operations
Full text linkProximity operations are becoming increasingly more important for current and future space missions, particularly
On-Orbit-Servicing (OOS) and Active-Debris Removal (ADR) ones. In this framework, a high-accuracy estimation
of the relative pose (position and attitude) between spacecraft is required to successfully and safely achieve complex
proximity operations like inspection, rendezvous, and docking. Visual navigation has recently become one of the most popular techniques for this purpose, thanks to the availability of increasingly compact, precise, and reliable monocular cameras. Traditional approaches relying on hand-engineered feature matching do not guarantee robustness or sufficient generalization, whereas Convolutional Neural Network (CNN)-based architectures have demonstrated improved robustness, noise rejection, and resilience to unseen scenarios. Despite their potential, these algorithms do not frequently reach the desired accuracy due, among others, to the employment of heuristic approaches in the choice of hyperparameters and the unavailability of an adequate large dataset.
This work proposes a CNN-based architecture for non-cooperative spacecraft monocular pose estimation exploiting optimization techniques to overcome these limits, improve performances and reduce the computational effort. This is achieved through the usage of a robust analytical method to select the best set of hyperparameters to minimize the pose loss function and the enhancement of the dataset for better feature learning. Moreover, the relationship between hyperparameters and the objective function (pose loss) is investigated, as well as the impact of different sets of hyperpa-
rameters on the CNN performance. A Blender® based synthetic dataset of approximately 25,000 synthetic images of an uncooperative target is generated to train the CNN. Such images are used to emulate representative proximity scenarios to validate the proposed approach.
The obtained results show that the proposed algorithm achieves centimeter-level position accuracy and near-degree-level attitude accuracy, maintaining, at the same time, high robustness against changes of illumination conditions and background textures
- …
