When dealing with proximity operations between a Target and a Chaser spacecraft, the design of an effective inspection strategy requires an approach that is able to consider many requirements and constraints. Therefore, the Chaser has to be able to evaluate in real time many possible solutions to identify and perform the optimal maneuver. The fulfillment of stringent safety and observation constraints, as well as the minimization of propellant consumption, represent the figures of merit to be optimized in order to consider the strategy successful in terms of inspection quality and costs. The nonlinearity of such requirements hinders the effectiveness of classical trajectory optimization methods. A method for the design of proximity maneuvers is to exploit the periodic natural motion trajectories through a graph search [8]. An alternative approach utilizes Sampling-based Motion Planning (SBMP) algorithms due to their effectiveness in highly complex environments, whereas the relative motion is computed using the linear solutions given by the Clohessy-Wiltshire (CW) model [14]. In this work an SBMP algorithm, based on Differential Algebra (DA) is proposed. Starting from an arbitrary command space, the algorithm explores, through a heuristic, the most relevant regions of the control space and then associates to each possible maneuver a specific score, which is related to the mission goals and constraints. To estimate the reward, the trajectory of the relative dynamics must also be quickly predicted. Thanks to the use of arbitrary orders Taylor polynomials, the DA technique allows to approximate the relative dynamics with improved accuracy with respect to CW, while still avoiding numerical integrations. Consequently, the proposed approach enables a trade-off between accuracy and computational efficiency. The performance of the proposed DA-Sampling based algorithm is assessed on a numerical test case compared to approaches from literature. The work shows that the proposed technique achieves the inspection goal while fulfilling safety requirements and minimizing cost. This is expected to open the possibility for a future autonomous on-board planning of the inspection maneuvers.
Quando si ha a che fare con operazioni di prossimità tra due satelliti, definiti Chaser e Target, la progettatione di una strategia di ispezione effettiva richiede un approccio capace di considerera molti requisiti e vincoli. Quindi, il Chaser deve essere capace di valutare in tempo reale molte solutioni per identificare ed effettuare la manovra ottimale. Il raggiungimento di vincoli stringenti di sicurezza ed osservazione, così come la minimizzazione di propellente, rappresentano figure di merito per considerare la strategia un successo in termini di costi e qualità. La nonlinearià mostrata da questi requisiti ostacola l'efficacia di metodi di ottimizzazione di traiettoria classici. Un metodo per la progettazione di manovre di prossimità sfrutta le traiettorie di moto naturale usando una ricerca grafica [8]. Un approccio alternativo utilizza un algoritmo SBMP (Sampling-Based Motion Planning) grazie alla sua efficacia in ambienti altamente complessi, e il modello Clohessy-Wiltshire (CW) [14] per descrivere il moto relativo. In questo lavoro un algoritmo SBMP, basato sull'algebra differenziale (DA) viene proposto. Partendo da un arbitrario spazio di comando campionato, l'algoritmo esplora, attraverso un'euristica, le regioni più rilevanti dello spazio di controllo e poi associa a ciascuna traiettoria un valore, legato agli obiettivi ed ai vincoli della missione. Per stimare questo valore, la traiettoria della dinamica relativa deve essere predetta velocemente. Grazie all'uso di polinomi di Taylor di ordine arbitrario, la tecnica DA permette di approssimare la dinamica relativa con accuratezza migliore rispetto a quella offerta da CW, senza dover effettuare integrazioni numeriche. Di conseguenza, l'approccio proposto offre un compromesso tra l'accuratezza e l'efficienza computazionale. Le prestazioni dell'algoritmo proposto vengono valutate con un test numerico confrontato con gli approcci della letteratura. Il lavoro mostra che la tecnica proposta raggiunge l'obiettivo della missione, rispettando i requisiti di sicurezza e costi. Questo apre alla possibilità di sviluppo di pianificazione autonoma di manovre di ispezione a bordo del Satellite.
Sampling-based approach for on-orbit satellite inspection
VITOLO, MARIO DANIELE
2018/2019
Abstract
When dealing with proximity operations between a Target and a Chaser spacecraft, the design of an effective inspection strategy requires an approach that is able to consider many requirements and constraints. Therefore, the Chaser has to be able to evaluate in real time many possible solutions to identify and perform the optimal maneuver. The fulfillment of stringent safety and observation constraints, as well as the minimization of propellant consumption, represent the figures of merit to be optimized in order to consider the strategy successful in terms of inspection quality and costs. The nonlinearity of such requirements hinders the effectiveness of classical trajectory optimization methods. A method for the design of proximity maneuvers is to exploit the periodic natural motion trajectories through a graph search [8]. An alternative approach utilizes Sampling-based Motion Planning (SBMP) algorithms due to their effectiveness in highly complex environments, whereas the relative motion is computed using the linear solutions given by the Clohessy-Wiltshire (CW) model [14]. In this work an SBMP algorithm, based on Differential Algebra (DA) is proposed. Starting from an arbitrary command space, the algorithm explores, through a heuristic, the most relevant regions of the control space and then associates to each possible maneuver a specific score, which is related to the mission goals and constraints. To estimate the reward, the trajectory of the relative dynamics must also be quickly predicted. Thanks to the use of arbitrary orders Taylor polynomials, the DA technique allows to approximate the relative dynamics with improved accuracy with respect to CW, while still avoiding numerical integrations. Consequently, the proposed approach enables a trade-off between accuracy and computational efficiency. The performance of the proposed DA-Sampling based algorithm is assessed on a numerical test case compared to approaches from literature. The work shows that the proposed technique achieves the inspection goal while fulfilling safety requirements and minimizing cost. This is expected to open the possibility for a future autonomous on-board planning of the inspection maneuvers.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_Mario_Daniele_Vitolo.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Tesi Mario Daniele Vitolo
Dimensione
4.62 MB
Formato
Adobe PDF
|
4.62 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/148369