The growing congestion of Earth’s orbital environment presents significant challenges for Space Situational Awareness and space objects catalogue maintenance. The rapid increase of large satellite constellations, such as Starlink, is increasing space traffic complexity, requiring more advanced solutions for tracking and cataloguing. The increasing involvement of military organizations, scientific missions, and commercial operators in space poses a major challenge for the tracking and surveillance community, as manoeuvre plans are rarely disclosed beforehand. This prevents space surveillance systems from effectively incorporating the manoeuvres of high-priority spacecrafts into trajectory predictions, making track correlation and orbit determination more complex and less reliable. This thesis presents several sequential estimation methods designed to enable the automatic, efficient and robust tracking of manoeuvring space objects during catalogue maintenance. An alternative is presented aiming at maintaining the trackability of the space object by characterising the uncertainty introduced by the manoeuvre, rather than the manoeuvre itself, using covariance inflation within sequential estimation. This may prevent us from inferring the manoeuvre characteristics, which is not required for catalogue maintenance purposes, while ensuring continuous tracking of the object. To determine the timing of the covariance inflation, two manoeuvre detection methods are implemented for redundancy. The first one explores the Mahalanobis distance of track attributables, using the predicted orbits and covariances within the so-called Manoeuvre Detection Filter built on an Extended Kalman Filter, while the second one makes use of a fixed-interval smoother in conjunction with the McReynolds consistency statistic. The methodologies are clearly detailed, along with test cases that evaluate their effectiveness and suitability. The results demonstrate that, by combining these different ideas, a sequential filter can be an interesting alternative for catalogue maintenance operations. Results are presented for simulated radar data, highlighting the performance of the detection and tracking strategy.
La crescente congestione dello spazio intorno alla Terra presenta sfide significative per la Space Situational Awareness e la manutenzione dei cataloghi di oggetti spaziali. Il rapido aumento delle costellazioni di satelliti sta incrementando la complessità del traffico spaziale, richiedendo soluzioni più avanzate per il tracking e la catalogazione. La crescente partecipazione di organizzazioni militari e operatori commerciali nello spazio rappresenta una sfida importante per la comunità di monitoraggio e sorveglianza, dato che i piani di manovra sono raramente resi noti in anticipo. Ciò impedisce ai sistemi di sorveglianza di incorporare efficacemente le manovre dei veicoli ad alta priorità nelle previsioni di traiettoria, rendendo la correlazione delle tracce e la determinazione dell’orbita più complessa e meno affidabile. Questa tesi presenta diversi metodi di stima sequenziale progettati per consentire il tracking automatico, efficiente e robusto di oggetti spaziali in manovra. Viene presentata un’alternativa che mira a mantenere la tracciabilità dell’oggetto spaziale caratterizzando l’incertezza introdotta dalla manovra, piuttosto che la manovra stessa, utilizzando l’inflazione della covarianza nell’ambito della stima sequenziale. In questo modo si può evitare di dedurre le caratteristiche della manovra, cosa non necessaria ai fini della manutenzione del catalogo, garantendo al contempo il tracking continuo dell’oggetto. Per determinare il momento dell’inflazione della covarianza, sono stati implementati due metodi di rilevamento delle manovre. Il primo esplora la distanza Mahalanobis degli attributables della traccia, utilizzando le orbite e le covarianze previste all’interno del cosiddetto Manoeuvre Detection Filter costruito su un Extended Kalman Filter, mentre il secondo fa uso di uno smoother a intervallo fisso in combinazione con la McReynolds consistency. Le metodologie sono dettagliate, insieme a casi di prova che ne valutano l’efficacia e l’idoneità. I risultati dimostrano che, combinando queste diverse idee, un filtro sequenziale può essere un’alternativa interessante per le operazioni di manutenzione del catalogo. I risultati sono presentati per dati radar simulati, evidenziando le prestazioni della strategia di rilevamento e tracking.
Sequential estimation methods for catalogue maintenance of manoeuvring objects
FLÓREZ GONZÁLEZ, ÁLVARO
2024/2025
Abstract
The growing congestion of Earth’s orbital environment presents significant challenges for Space Situational Awareness and space objects catalogue maintenance. The rapid increase of large satellite constellations, such as Starlink, is increasing space traffic complexity, requiring more advanced solutions for tracking and cataloguing. The increasing involvement of military organizations, scientific missions, and commercial operators in space poses a major challenge for the tracking and surveillance community, as manoeuvre plans are rarely disclosed beforehand. This prevents space surveillance systems from effectively incorporating the manoeuvres of high-priority spacecrafts into trajectory predictions, making track correlation and orbit determination more complex and less reliable. This thesis presents several sequential estimation methods designed to enable the automatic, efficient and robust tracking of manoeuvring space objects during catalogue maintenance. An alternative is presented aiming at maintaining the trackability of the space object by characterising the uncertainty introduced by the manoeuvre, rather than the manoeuvre itself, using covariance inflation within sequential estimation. This may prevent us from inferring the manoeuvre characteristics, which is not required for catalogue maintenance purposes, while ensuring continuous tracking of the object. To determine the timing of the covariance inflation, two manoeuvre detection methods are implemented for redundancy. The first one explores the Mahalanobis distance of track attributables, using the predicted orbits and covariances within the so-called Manoeuvre Detection Filter built on an Extended Kalman Filter, while the second one makes use of a fixed-interval smoother in conjunction with the McReynolds consistency statistic. The methodologies are clearly detailed, along with test cases that evaluate their effectiveness and suitability. The results demonstrate that, by combining these different ideas, a sequential filter can be an interesting alternative for catalogue maintenance operations. Results are presented for simulated radar data, highlighting the performance of the detection and tracking strategy.File | Dimensione | Formato | |
---|---|---|---|
2025_07_Florez_Gonzalez.pdf
accessibile in internet per tutti
Descrizione: Master thesis report
Dimensione
7.68 MB
Formato
Adobe PDF
|
7.68 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/10589/239977