1,721,254 research outputs found
A geometrical approach for the angular velocity determination using a star sensor
No full textIn this paper, a geometrical investigation of the star sensor image is performed under dynamic conditions where the angular velocity effects are non-negligible. It is shown that, when the spacecraft is rotating, the streaks left by the stars’ signal onto the star sensor detector belong to portions of conic sections which features depend on the angles between the instantaneous rotation axis, the sensor line of sight and the stars’ direction. The geometrical properties discussed in the first part of the paper can be used to develop new numerical methods for the evaluation of the angular velocity. Hence, the chord method is proposed and discussed. This approach needs at least two stars in two successive images and, despite its simplicity, is quite effective to get a preliminary estimation of the spacecraft angular velocity in terms of direction and magnitude. Using the stars’ centroids from two successive images, the chord method evaluates the angular velocity direction as the intersection of the normals to the streaks. The proposed method is firstly presented by means of simple examples using some reference geometries, and then it is applied to real scenarios by using a high fidelity star sensor simulator. The pre-processing and processing of simulated images are discussed, presenting the geometrical techniques used to cluster the streaks and compute the centroids. Results are presented and discussed, validating the reported theoretical speculations
Inverse-dynamics particle swarm optimization for real time optimal control: Challenges and opportunities
No full textThe paper described the most recent advances in the development of the Inverse-dynamics Particle Swarm Optimization technique for determining approximate solutions to constrained optimal control problems. Near-optimal solutions are searched for using a differential flatness approach such that the kinematic of the problem is directly approximated whereas the dynamics is given by a generic non-linear combination of successive time-derivatives of the state. Numerical results are presented evaluating the proposed technique with a constrained reorientation maneuver of a spacecraft already considered in literature. A rest-to-rest slew maneuver is considered where an optical sensor cannot be exposed to sources of bright light such as the Earth, the Sun and the Moon. It is established that the computation of minimum time, minimum fuel and minimum energy maneuvers with the proposed technique leads to near optimal solutions, which fully satisfy all the boundary and path constraints. The rapid convergence characterizes the proposed technique as a feasible future on-board path-planner for terrestrial and space applications
Inverse dynamics particle swarm optimization applied to constrained minimum-time maneuvers using reaction wheels
No full textThe paper deals with the problem of time-optimal spacecraft reorientation maneuvers by means of reaction wheels, with boundary and path constraints. When searching for solutions to optimal attitude-control problems, spacecraft can be easily modeled as controlled by external torques. However, when using actuators such as reaction wheels, conservation of the total angular momentum must be taken into account and the wheel dynamics must be included. A rest-to-rest slew maneuver is considered where an optical sensor cannot be exposed to sources of bright light such as the Earth, the Sun and the Moon. The motion must be constrained to prevent the sensor axis from entering into established keep-out cones. The minimum-time solution is proposed using the Inverse Dynamics Particle Swarm Optimization technique. The attitude and the kinematics of the satellite evolve, leading to the successive attainment of the wheel control input via fixed-step numerical integration. Numerical results are evaluated over different scenarios. It is established that the computation of minimum time maneuvers with the proposed technique leads to near optimal solutions, which fully satisfy all the boundary and path constraints. The ability to converge in a variety of different scenarios always requiring the same computational effort characterizes the proposed technique as a feasible future on-board path-planner
Design and Testing of a Demonstrator Electric-Pump Feed System for Liquid Propellant Rocket Engines
No full textThe construction of an experimental test rig of an electric-pump feed system for liquid-propellant rocket engines is discussed, focusing in particular on the choice of the different components, and the evaluation of their suitability. Most components are chosen off-the-shelf, yet they are proven to result in an efficient, lightweight design. However, the propellant pumps are found to require instead a customized design for actual inflight applications. Unlike other components, the injection plate is designed and machined ad hoc for the present test bench. It features a single pentad injector to accomodate for the widely different flowrates of oxidizer and fuel. Hydraulic oil is used instead of the actual propellants due to restrictions deriving from adopting off-the-shelf pumps, which however is likely to entail limitations not greater than those associated with standard water testing of liquid-propellant rocket injection plates. The five jets are found to intersect positively. The characteristic curves and relationships of the oxidizer and fuel hydraulic circuits are also shown and discussed
Particle swarm with domain partition and control assignment for time-optimal maneuvers
No full textA novel approach has been proposed for planning time-optimal maneuvers imposing a bang–bang external control. The optimizer was based on the particle swarm optimization and only required setting the maximum number of switches allowed for each axis. Two different test cases were analyzed and solved to validate the optimizer. In the first example, characterized by four state-space variables and no path constraints, the convergence toward the optimal solution has been demonstrated with different values of the maximum number of switches. For the second example, described by six state-space variables and nonlinear path constraints, the optimal solution was reached in more than 70% of the cases over 100 simulations, path constraints were completely satisfied, and errors on the final conditions were lower than 10-8 in normalized units. The computational effort was rather small because about 2000 iterations were required to reach convergence. Moreover, the second example has shown that the solution obtained with the described method could be better than the solutions obtained with pseudospectral optimization algorithms
Autonomous crater detection on asteroids using a fully-convolutional neural network
No full textThis paper shows the application of autonomous Crater Detection using a Fully-Convolutional Neural Network (FCN) on a main belt asteroid, Ceres. The FCN, namely a U-Net, is trained on optical images of the Moon Global Morphology Mosaic based on data collected by the Lunar Recoinnassance Orbiter (LRO), using manual crater catalogues as annotations for the creation of ground truth data. The Moon-trained network will then be tested on optical images of Ceres taken during the Dawn mission; as this network is adapted to a different domain from the one upon which it was already trained, this task is accomplished by means of a Transfer Learning (TL) approach. In particular, this work will take into consideration a homogeneous TL, because the image features and labels share the same qualities in both domains. The Moon-trained model has been fine-tuned using 100, 500 and 1000 additional annotated images of Ceres, reaching a testing accuracy on 350 never before seen images of 96.24%, 96.95% and 97.19%, respectively. Annotations for Ceres are taken from the Zeilnhofer crater catalogue, containing 44.594 craters with diameter greater than 1 km in a latitude range of 84.66°S-89.62°N, full longitude: therefore, a near-global coverage of the asteroid surface will be taken into consideration.
The output of the U-Net is a grey-scale mask containing predicted craters: it will be post-processed applying global thresholding for image binarization and a template matching algorithm to extract craters positions and radii in the pixel space: this allows the creation of a definitive binary mask. These post-processed craters will be counted and compared to the ground truth data in order to compute image segmentation metrics: precision, recall and their harmonic mean known as F1 score. Precision, which is the ratio between matched craters and all the craters found by the network, reached values of 70.51%, 84.43% and 83.45%. Recall, defined as the ratio between matched and catalogued craters, achieved performances of 48.27%, 53.32% and 58.97%; the lower values are due to the fact that when precision is high, recall is automatically penalized and vice versa. The F1 score is introduced to sum up the effects of both the above indices: its reached percentages are 57.30%, 65.36% and 69.11%. These results are considered to be encouraging because they are comparable with other works that use the U-Net in a Crater Detection contex
Improved magnetic charged system search optimization algorithm with application to satellite formation flying
Full text linkThis paper is devoted to the implementation and application of an improved version of the metaheuristic algorithm called magnetic charged system search. Some modifications and novelties are introduced and tested. Firstly, the authors’ attempt is to develop a self-adaptive and user-friendly algorithm which can automatically set all the preliminary parameters (such as the numbers of particles, the maximum iterations number) and the internal coefficients. Indeed, some mathematical laws are proposed to set the parameters and many coefficients can dynamically change during the optimization process based on the verification of internal conditions. Secondly, some strategies are suggested to enhance the performances of the proposed algorithm. A chaotic local search is introduced to improve the global best particle of each iteration by exploiting the features of ergodicity and randomness. Moreover, a novel technique is proposed to handle bad-defined boundaries; in fact, the possibility to self-enlarge the boundaries of the optimization variables is considered, allowing to achieve the global optimum even if it is located on the boundaries or outside. The algorithm is tested through some benchmark functions and engineering design problems, showing good results, followed by an application regarding the problem of time-suboptimal manoeuvres for satellite formation flying, where the inverse dynamics technique, together with the B-splines, is employed. This analysis proves the ability of the proposed algorithm to optimize control problems related to space engineering, obtaining better results with respect to more common and used algorithms in literature
Time suboptimal formation flying manoeuvres through improved magnetic charged system search
Full text linkThe development of fast and reliable optimization algorithms is required in order to obtain real-time optimal trajectory on-board spacecraft. In addition, the wide spread of small satellites, due to their low costs, is leading to a greater number of satellite formations in space. This paper presents an Improved version of the Magnetic Charged System Search (IMCSS) metaheuristic algorithm to compute time-suboptimal manoeuvres for satellite formation flying. The proposed algorithm exploits some strategies aimed at improving the convergence to the optimum, such as the chaotic local search and the boundary handling technique, and it is able to self-tune its internal parameters and coefficients. Moreover, the inverse dynamics technique and the differential flatness approach, through the B-splines curves, are used to approximate the trajectory. The optimization procedure is applied to the circular J2 relative model developed by Schweighart and Sedwick and to the elliptical relative motion model developed by Yamanaka and Ankersen. The results of this paper show that the convergence is better achieved by using the proposed tools, thus proving the efficiency and reliability of the algorithm in solving some space engineering problems
- …
