This project, developed at the Surgical Robotics Lab, University of Twente, The Netherlands, is part of a vast field of research and development that is `Minimally Invasive Surgery' and BigMag is a novel device that falls into this field because its purpose is to develop technology to deflect the tips of endovascular catheters using magnetic fields. Endovascular catheters belong to continuum manipulator class, therefore it is necessary to derive models to describe their behavior from data collected online by tracking systems. This project aimed at developing an advanced model-based tracking algorithm in order to improve the techniques used for BigMag. The main purpose of the algorithm was to obtain information on the catheter's configuration used for BigMag system with high precision. Two algorithms have been developed. The first model-based algorithm has been developed with a linear fitting technique to describe the catheter as a series of rigid links connected by joints. The algorithm has been validated, both with stereo vision data provided by BigMag and synthetic data generated with CAD software, in the reconstruction of configuration parameters of the catheter. The second algorithm has been developed exploiting the PRBM technique, allowing to implement the relationship between the magnetic field provided by BigMag and the final configuration of the catheter, including its elastic characteristics. The model has been enhanced with different techniques, introducing a new compliance model and a Luenberger State Observer. PRBM algorithm has been implemented on Simulink in order to perform the evaluation of the performance and validation in comparison with the linear fitting algorithm implemented on BigMag tracking system.The linear fitting algorithm showed a behavior comparable with literature with sample mean and standard deviation of the relative error of (1.2,0.86%) in the reconstruction of the radius of curvature with real data and (2.39,1.64%) exploiting a synthetic dataset. The final results of PRBM algorithm showed satisfactory results in the evaluation of the tip's position of (-0.007,0.11cm), much better than initial evaluations (0.51,0.68cm).
Questo progetto, sviluppato al Surgical Robotics Lab, Università di Twente, fa parte di un campo di ricerca e denominato `Chirurgia Mini-invasiva' e BigMag è un nuovo dispositivo che rientra in questo campo poichè sviluppa tecnologie di deflessione magnetica di cateteri intravascolari per fini medici. Questi cateteri fanno parte di una classe di dispositivi detta`manipolatori continui', per cui è necessario elaborare modelli per descriverne il comportamento a partire dai dati raccolti dai sistemi di tracking. Questo progetto mirava a sviluppare un algoritmo di model-based tracking per migliorare le tecniche presenti su BigMag. Lo scopo principale dell'algoritmo era quello di ottenere informazioni sulla configurazione dei cateteri utilizzati dal sistema con elevata precisione. Due algoritmi sono stati sviluppati e validati. Il primo algoritmo di model-based tracking è stato sviluppato con una tecnica di fitting lineare descrivendo il catetere come una serie di segmenti rigidi connessi da giunti. L'algoritmo è stato validato, sia con dati forniti dal sistema ottico di BigMag sia da dati sintetici generati da software CAD, nella valutazione della configurazione del catetere. Il secondo algoritmo è stato sviluppato con la tecnica PRBM, introducendo la relazione fra il campo magnetico di BigMag, la configurazione del catetere e le sue proprietà elastiche. Il modello è stato migliorato con più tecniche fra cui l'introduzione di un nuovo modello di compliance e l’osservatore di Luenberger. L'algoritmo PRBM è stato implementato su Simulink per valutare la performance e la validazione, paragonando i risultati con l'algoritmo di fitting lineare implementato sul sistema di tracking di BigMag. Il primo algoritmo di fitting lineare ha mostrato un comportamento paragonabile alla letteratura con una media campionaria e deviazione standard delll'errore pari a (1.2,0.6%) nella ricostruzione del raggio di curvatura con dati reali e (2.39,1.64%) sfruttando i dati sintetici I risultati finali dell'algoritmo PRBM hanno mostrato risultati soddisfacenti nella valutazione della posizione della punta del catetere (-0.007,0.11cm), migliori rispetto ai primi test (0.51,0.68cm).
Enhanced fitting techniques for model-based tracking of continuum manipulators
BUCCHI, GIUSEPPE
2017/2018
Abstract
This project, developed at the Surgical Robotics Lab, University of Twente, The Netherlands, is part of a vast field of research and development that is `Minimally Invasive Surgery' and BigMag is a novel device that falls into this field because its purpose is to develop technology to deflect the tips of endovascular catheters using magnetic fields. Endovascular catheters belong to continuum manipulator class, therefore it is necessary to derive models to describe their behavior from data collected online by tracking systems. This project aimed at developing an advanced model-based tracking algorithm in order to improve the techniques used for BigMag. The main purpose of the algorithm was to obtain information on the catheter's configuration used for BigMag system with high precision. Two algorithms have been developed. The first model-based algorithm has been developed with a linear fitting technique to describe the catheter as a series of rigid links connected by joints. The algorithm has been validated, both with stereo vision data provided by BigMag and synthetic data generated with CAD software, in the reconstruction of configuration parameters of the catheter. The second algorithm has been developed exploiting the PRBM technique, allowing to implement the relationship between the magnetic field provided by BigMag and the final configuration of the catheter, including its elastic characteristics. The model has been enhanced with different techniques, introducing a new compliance model and a Luenberger State Observer. PRBM algorithm has been implemented on Simulink in order to perform the evaluation of the performance and validation in comparison with the linear fitting algorithm implemented on BigMag tracking system.The linear fitting algorithm showed a behavior comparable with literature with sample mean and standard deviation of the relative error of (1.2,0.86%) in the reconstruction of the radius of curvature with real data and (2.39,1.64%) exploiting a synthetic dataset. The final results of PRBM algorithm showed satisfactory results in the evaluation of the tip's position of (-0.007,0.11cm), much better than initial evaluations (0.51,0.68cm).| File | Dimensione | Formato | |
|---|---|---|---|
|
2018_4_Bucchi.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
22.46 MB
Formato
Adobe PDF
|
22.46 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/10589/140235