Flying humanoid robots represent a cutting-edge intersection of aerial mobility and humanoid design, offering unprecedented capabilities for navigation and interaction in diverse environments. This thesis delves into the intricate challenges and advancements in the development of flying humanoid robots, with a specific focus on thrust estimation for stable and controlled flight. Leveraging insights from flying robotics, humanoid robotics, and external wrench estimation, the study investigates various approaches to integrating sensor data with mathematical models to accurately estimate thrust intensities for jet-powered humanoid robots. The thesis sets the stage by providing an overview of humanoid robots' capabilities and introduces the iRonCub project as a pioneering effort in this field. It poses several key research questions, addressing the applicability of existing models for thrust estimation and the potential use of sensors commonly found in humanoid robots. Further, the thesis delves into a comprehensive review of the state-of-the-art, exploring existing jet engine-based thrust estimation models and sensor-based approaches. Experimental campaigns are conducted to assess the feasibility of leveraging force/torque sensors typically available in humanoid robots for thrust estimation. The study highlights the limitations of existing models and the challenges of in-situ sensor calibration. Innovative approaches are proposed, including the adoption of centroidal momentum-based estimation and the combination of multiple estimation methods into a unified framework. Simulation and real-world validation experiments are conducted to evaluate the performance and robustness of these approaches. Despite promising results, challenges such as the lack of ground-truth validation and the limitations of individual methods are acknowledged. In the end, the thesis reflects on the attempted solutions to the main research question and acknowledges the inherent trade-offs and complexities involved in thrust estimation for flying humanoid robots. It emphasizes the importance of combining multiple approaches to achieve more robust and accurate estimations. In conclusion, this thesis contributes to advancing the understanding and capabilities of flying humanoid robots, shedding light on the challenges and opportunities in thrust estimation. Future work is proposed to further refine sensor calibration procedures, enhance estimation algorithms, and explore interdisciplinary approaches to address the complexities of flying humanoid robotics. Through continued research and innovation, flying humanoid robots hold the potential to revolutionize various fields, from search and rescue to environmental monitoring and beyond.
I robot umanoidi volanti rappresentano un'avanguardia tecnologica che combina mobilità aerea e design umanoide, offrendo capacità senza precedenti per la navigazione e l'interazione in ambienti diversi. Questa tesi affronta le sfide complesse e i progressi nello sviluppo di robot umanoidi volanti, con un focus specifico sulla stima della spinta per un volo stabile e controllato. Sfruttando conoscenze nell'ambito della robotica volante, della robotica umanoide e della stima delle forze esterne, lo studio esplora diversi approcci per integrare i dati dei sensori con modelli matematici al fine di stimare accuratamente l'intensità della spinta per robot umanoidi a propulsione jet. La tesi introduce le capacità dei robot umanoidi e presenta il progetto iRonCub come un'iniziativa pionieristica in questo campo. Vengono poste alcune domande chiave di ricerca, affrontando l'applicabilità dei modelli esistenti per la stima della spinta e il potenziale utilizzo di sensori comunemente integrati nei robot umanoidi. Inoltre, viene svolta una revisione completa dello stato dell'arte, analizzando i modelli di stima della spinta basati su motori jet e approcci fondati sui sensori. Sono condotte campagne sperimentali per valutare la fattibilità dell'impiego di sensori di forza/coppia tipicamente disponibili nei robot umanoidi per la stima della spinta. Lo studio evidenzia i limiti dei modelli esistenti e le sfide legate alla calibrazione dei sensori in situ. Vengono proposti approcci innovativi, tra cui l'adozione della stima basata sul momento centroidale e la combinazione di diversi metodi di stima in un unico framework. Esperimenti di simulazione e validazione nel mondo reale vengono condotti per valutare le prestazioni e la robustezza di questi approcci. Nonostante i risultati promettenti, vengono riconosciute sfide come l'assenza di una validazione di riferimento e i limiti dei singoli metodi. Infine, la tesi riflette sulle soluzioni tentate per rispondere alla principale domanda di ricerca, riconoscendo i compromessi e le complessità intrinseche nella stima della spinta per i robot umanoidi volanti. Viene sottolineata l'importanza di combinare più approcci per ottenere stime più robuste e accurate. In conclusione, questa tesi contribuisce a far progredire la comprensione e le capacità dei robot umanoidi volanti, evidenziando le sfide e le opportunità legate alla stima della spinta. Si propongono sviluppi futuri per perfezionare le procedure di calibrazione dei sensori, migliorare gli algoritmi di stima e esplorare approcci interdisciplinari per affrontare le complessità della robotica umanoide volante. Attraverso la ricerca e l'innovazione continue, i robot umanoidi volanti hanno il potenziale di rivoluzionare diversi settori, dal soccorso in situazioni di emergenza al monitoraggio ambientale e oltre.
Nonlinear estimation of force fields for aerial humanoid robots
Mohamed, Hosameldin Awadalla Omer
2024/2025
Abstract
Flying humanoid robots represent a cutting-edge intersection of aerial mobility and humanoid design, offering unprecedented capabilities for navigation and interaction in diverse environments. This thesis delves into the intricate challenges and advancements in the development of flying humanoid robots, with a specific focus on thrust estimation for stable and controlled flight. Leveraging insights from flying robotics, humanoid robotics, and external wrench estimation, the study investigates various approaches to integrating sensor data with mathematical models to accurately estimate thrust intensities for jet-powered humanoid robots. The thesis sets the stage by providing an overview of humanoid robots' capabilities and introduces the iRonCub project as a pioneering effort in this field. It poses several key research questions, addressing the applicability of existing models for thrust estimation and the potential use of sensors commonly found in humanoid robots. Further, the thesis delves into a comprehensive review of the state-of-the-art, exploring existing jet engine-based thrust estimation models and sensor-based approaches. Experimental campaigns are conducted to assess the feasibility of leveraging force/torque sensors typically available in humanoid robots for thrust estimation. The study highlights the limitations of existing models and the challenges of in-situ sensor calibration. Innovative approaches are proposed, including the adoption of centroidal momentum-based estimation and the combination of multiple estimation methods into a unified framework. Simulation and real-world validation experiments are conducted to evaluate the performance and robustness of these approaches. Despite promising results, challenges such as the lack of ground-truth validation and the limitations of individual methods are acknowledged. In the end, the thesis reflects on the attempted solutions to the main research question and acknowledges the inherent trade-offs and complexities involved in thrust estimation for flying humanoid robots. It emphasizes the importance of combining multiple approaches to achieve more robust and accurate estimations. In conclusion, this thesis contributes to advancing the understanding and capabilities of flying humanoid robots, shedding light on the challenges and opportunities in thrust estimation. Future work is proposed to further refine sensor calibration procedures, enhance estimation algorithms, and explore interdisciplinary approaches to address the complexities of flying humanoid robotics. Through continued research and innovation, flying humanoid robots hold the potential to revolutionize various fields, from search and rescue to environmental monitoring and beyond.File | Dimensione | Formato | |
---|---|---|---|
2024_12_Thesis_Mohamed.pdf
accessibile in internet solo dagli utenti autorizzati
Dimensione
46.19 MB
Formato
Adobe PDF
|
46.19 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/10589/232032