In recent years, drones have increasingly been utilized in both civilian and military contexts for operations such as search and rescue, aerial mapping, disaster management, surveillance, monitoring, and recreational light shows. Research has shifted towards drone swarms, focusing on integrating various levels of autonomy. Typically, these operations involve one-to-many scenarios, where a single operator manages multiple drones. Starting from an analysis of the possible requirements, this master’s thesis aims to develop and validate a system architecture for managing and simulating multiple drones within a Software-in-the-Loop (SITL) environment. An examination of european drone legislation and potential use cases was conducted to derive the necessary requirements and functionalities. The developed architecture includes a simulation environment created through Unreal Engine, using Cesium plugin for the ground model and AirSim plugin for the aero-mechanical model of the drone. As the flight controller, the open-source PX4-Autopilot plugin was used. A ground station prototype was built in Python and includes a mission management command module and a control module to gather telemetry data, communicating with the drones through the Micro Aerial Vehicle Link (MAVLink) protocol. Validation of the system was performed through simulations related to the aerial mapping use case supported by a cooperative coverage path planning algorithm, i.e. multiple drones will have to scan a single area to optimize mission time. The path planning algorithm is not within the scope of this thesis, as the work was conducted during a company internship involving a larger team at TXT E-TECH s.r.l. The results obtained meet those expected, having an architecture that is not only func tional for the chosen use case but also mission-type independent. It is scalable in terms of drone numbers within PX4 limitations, and ensures positional accuracy through both the MAVLink protocol and the Cesium integration in Unreal. Moreover, several mission management processes have been automated. Keywords: MVMS, Drones, Requirements, Architecture, Mission, Ground station, Simulation, MAVLink
Negli ultimi anni, i droni sono stati sempre più utilizzati sia in contesti civili che militari per operazioni come la ricerca e il soccorso, la mappatura aerea, la gestione dei disastri, la sorveglianza, il monitoraggio e gli spettacoli luminosi ricreativi. La ricerca si è spostata verso gli sciami di droni, concentrandosi sull’integrazione di vari livelli di autonomia. Tipicamente, queste operazioni coinvolgono scenari one-to-many, dove un singolo operatore gestisce più droni. Partendo da un’analisi dei possibili requisiti, questa tesi magistrale mira a sviluppare e validare un’architettura di sistema per la gestione e la simulazione di più droni all’interno di un ambiente Software-in-the-Loop (SITL). È stata condotta un’analisi della legislazione europea sui droni e dei possibili casi d’uso per derivare i requisiti e le funzionalità necessarie. L’architettura sviluppata include un ambiente di simulazione creato con Unreal Engine, utilizzando il plugin Cesium per il modello terrestre e il plugin AirSim per il modello aero-meccanico del drone. Come controllore di volo è stato utilizzato il plugin open-source PX4-Autopilot. È stato costruito un prototipo di ground station in Python che include un modulo di gestione della missione e un modulo di controllo per raccogliere dati di telemetria, comunicando con i droni tramite il protocollo Micro Aerial Vehicle Link (MAVLink). La validazione del sistema è stata eseguita attraverso simulazioni relative al caso d’uso della mappatura aerea, supportato da un algoritmo di cooperative coverage path planning per cui più droni dovranno scansionare un’area unica per ottimizzare il tempo di missione. L’algoritmo in questione non rientra nello scopo di questa tesi, poiché il lavoro è stato svolto in team più ampio durante uno stage aziendale presso TXT E-TECH s.r.l. I risultati ottenuti soddisfano le aspettative, con un’architettura non solo funzionale per il caso d’uso scelto, ma anche indipendente dal tipo di missione. È scalabile in termini di numero di droni entro i limiti imposti da PX4, e garantisce un’accuratezza posizionale grazie sia al protocollo MAVLink che all’integrazione di Cesium in Unreal. Inoltre, diversi processi di gestione delle missioni sono stati automatizzati. Parole chiave: MVMS, Droni, Requisiti, Architettura, Missione, Ground station, Simulazione, MAVLink
Requirements analysis and architecture validation of a Multi Vehicle Management System (MVMS) for drones
TOTH, ALBERTO
2023/2024
Abstract
In recent years, drones have increasingly been utilized in both civilian and military contexts for operations such as search and rescue, aerial mapping, disaster management, surveillance, monitoring, and recreational light shows. Research has shifted towards drone swarms, focusing on integrating various levels of autonomy. Typically, these operations involve one-to-many scenarios, where a single operator manages multiple drones. Starting from an analysis of the possible requirements, this master’s thesis aims to develop and validate a system architecture for managing and simulating multiple drones within a Software-in-the-Loop (SITL) environment. An examination of european drone legislation and potential use cases was conducted to derive the necessary requirements and functionalities. The developed architecture includes a simulation environment created through Unreal Engine, using Cesium plugin for the ground model and AirSim plugin for the aero-mechanical model of the drone. As the flight controller, the open-source PX4-Autopilot plugin was used. A ground station prototype was built in Python and includes a mission management command module and a control module to gather telemetry data, communicating with the drones through the Micro Aerial Vehicle Link (MAVLink) protocol. Validation of the system was performed through simulations related to the aerial mapping use case supported by a cooperative coverage path planning algorithm, i.e. multiple drones will have to scan a single area to optimize mission time. The path planning algorithm is not within the scope of this thesis, as the work was conducted during a company internship involving a larger team at TXT E-TECH s.r.l. The results obtained meet those expected, having an architecture that is not only func tional for the chosen use case but also mission-type independent. It is scalable in terms of drone numbers within PX4 limitations, and ensures positional accuracy through both the MAVLink protocol and the Cesium integration in Unreal. Moreover, several mission management processes have been automated. Keywords: MVMS, Drones, Requirements, Architecture, Mission, Ground station, Simulation, MAVLinkFile | Dimensione | Formato | |
---|---|---|---|
Toth_Alberto.pdf
accesso riservato
Dimensione
37.78 MB
Formato
Adobe PDF
|
37.78 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/74801