This thesis presents the development and validation of a novel collision avoidance algorithm, named the Avoidance of Collision Point Guidance (ACPG) algorithm, designed for fixed-wing unmanned aerial vehicle (UAV) swarms. Addressing the complexities and dynamic challenges inherent in UAV swarm operations, the ACPG algorithm innovatively synthesizes elements from sense and avoid, geometric, and vector (force or potential) field approaches. By leveraging the capabilities of Lidar technology, the algorithm efficiently detects obstacles, assessing their size, velocity, and distance, to dynamically calculate the necessary maneuvers for collision avoidance. This involves an inventive application of the inverted Proportional Navigation Guidance (PNG) technique, combined with geometric considerations, to generate a repulsive velocity vector field that guides both the leader UAV and the entire swarm. Extensive simulations demonstrate the algorithm’s effectiveness in 3D environments, showcasing its adaptability to both static and dynamic obstacles through minimal effort directional adjustments and simultaneous altitude modifications. The robustness of the ACPG algorithm is further evidenced through a series of case studies and sensitivity analyses concerning key design parameters (Nv and Nh), obstacle size variations, and reduced detection ranges. Despite these challenges, the algorithm consistently maintains optimal performance, ensuring the safety and operational integrity of the UAV swarm. In addition, a two-dimensional vector field avoidance algorithm is developed and tested via the simulations. This algorithm is specifically tailored to address the challenge of navigating around predefined, comparatively larger areas that pose navigational restrictions, such as no-fly zones. Unlike the dynamic and potentially smaller flying obstacles, these areas represent static obstacles that require a different strategic approach due to their extensive spatial coverage.
Questa tesi presenta lo sviluppo e la validazione di un nuovo algoritmo per evitare le collisioni, denominato algoritmo Evitamento del Punto di Collisione Guidance (ACPG), progettato per sciami di veicoli aerei senza pilota ad ala fissa (UAV). Affrontando le complessità e le sfide dinamiche inerenti alle operazioni di sciame di UAV, l’algoritmo ACPG sintetizza in modo innovativo elementi provenienti da approcci di campo di rilevamento ed evitamento, geometrici e vettoriali (forza o potenziale). Sfruttando le capacità della tecnologia Lidar, l’algoritmo rileva in modo efficiente gli ostacoli, valutandone dimensioni, velocità e distanza, per calcolare dinamicamente le manovre necessarie per evitare la collisione. Ciò comporta un’applicazione inventiva della tecnica di guida proporzionale alla navigazione invertita (PNG), combinata con considerazioni geometriche, per generare un campo vettoriale di velocità repulsiva che guida sia l’UAV leader che l’intero sciame. Ampie simulazioni dimostrano l’efficacia dell’algoritmo in ambienti 3D, mostrando la sua adattabilità sia agli ostacoli statici che dinamici attraverso regolazioni direzionali con sforzo minimo e modifiche simultanee dell’altitudine. La robustezza dell’algoritmo ACPG è ulteriormente evidenziata attraverso una serie di casi di studio e analisi di sensibilità riguardanti i parametri chiave di progettazione (Nv e Nh), variazioni delle dimensioni degli ostacoli e intervalli di rilevamento ridotti. Nonostante queste sfide, l’algoritmo mantiene costantemente prestazioni ottimali, garantendo la sicurezza e l’integrità operativa dello sciame di UAV. Inoltre, tramite le simulazioni viene sviluppato e testato un algoritmo bidimensionale per evitare il campo vettoriale. Questo algoritmo è specificamente studiato per affrontare la sfida della navigazione in aree predefinite e relativamente più grandi che pongono restrizioni alla navigazione, come le zone interdette al volo. A differenza degli ostacoli volanti dinamici e potenzialmente più piccoli, queste aree rappresentano ostacoli statici che richiedono un approccio strategico diverso a causa della loro ampia copertura spaziale.
Fixed-Wing UAV swarm obstacle avoidance using inverted proportional navigation guidance
USLU, FURKAN
2022/2023
Abstract
This thesis presents the development and validation of a novel collision avoidance algorithm, named the Avoidance of Collision Point Guidance (ACPG) algorithm, designed for fixed-wing unmanned aerial vehicle (UAV) swarms. Addressing the complexities and dynamic challenges inherent in UAV swarm operations, the ACPG algorithm innovatively synthesizes elements from sense and avoid, geometric, and vector (force or potential) field approaches. By leveraging the capabilities of Lidar technology, the algorithm efficiently detects obstacles, assessing their size, velocity, and distance, to dynamically calculate the necessary maneuvers for collision avoidance. This involves an inventive application of the inverted Proportional Navigation Guidance (PNG) technique, combined with geometric considerations, to generate a repulsive velocity vector field that guides both the leader UAV and the entire swarm. Extensive simulations demonstrate the algorithm’s effectiveness in 3D environments, showcasing its adaptability to both static and dynamic obstacles through minimal effort directional adjustments and simultaneous altitude modifications. The robustness of the ACPG algorithm is further evidenced through a series of case studies and sensitivity analyses concerning key design parameters (Nv and Nh), obstacle size variations, and reduced detection ranges. Despite these challenges, the algorithm consistently maintains optimal performance, ensuring the safety and operational integrity of the UAV swarm. In addition, a two-dimensional vector field avoidance algorithm is developed and tested via the simulations. This algorithm is specifically tailored to address the challenge of navigating around predefined, comparatively larger areas that pose navigational restrictions, such as no-fly zones. Unlike the dynamic and potentially smaller flying obstacles, these areas represent static obstacles that require a different strategic approach due to their extensive spatial coverage.File | Dimensione | Formato | |
---|---|---|---|
2024_04_Uslu_Thesis_01.pdf
accessibile in internet per tutti
Descrizione: Text of the thesis
Dimensione
8.51 MB
Formato
Adobe PDF
|
8.51 MB | Adobe PDF | Visualizza/Apri |
2024_04_Uslu_Executive_Summary_02.pdf
accessibile in internet per tutti
Descrizione: Executive summary
Dimensione
969.41 kB
Formato
Adobe PDF
|
969.41 kB | 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/10589/219379