University of Twente Student Theses

Login

Positioning and orientation control of a needle-insertion MRI compatible medical robot based on ROS using visual feedback

Shametaj, P. (2021) Positioning and orientation control of a needle-insertion MRI compatible medical robot based on ROS using visual feedback.

This is the latest version of this item.

[img] PDF
9MB
Abstract:Within the discipline of surgical interventions, there has been a focus on percutaneous nee-dle insertions. The procedure usually involves the insertion of a needle or probe deep inside the patient’s skin. These minimally invasive procedures are usually used for ablation, abscess drainages, biopsies or preventive purposes. The robot used in this project has been designed to be used as a patient mounted robot for irreversible electroporation (IRE) targeting pancreatic adenocarcinoma tumor cells. The project’s scope is creating a functional first prototype that could be used for the insertion of electrodes during the IRE procedure, with a focus on creating a software framework that can be adapted to future changes in the physical robot design. This thesis will skew from a traditional approach of a pure research or design thesis project and be a combination of two phases, Implementation and Design. This is due to the initial state of the project, where only the physical structure of the robot was available. In the first phase, an initial testing prototype upon which research could be performed was developed. With the use of ROS, a modular framework that integrated all the hardware components, calculated the robot kinematics and handled the communication within the software was designed. In the second phase, the project’s focus changed to researching a specific design decision, that of the robot feedback. After analysing the available options, the decision of using non-MRI visual feedback was made. In this stage, the accuracy and reliability that vision feedback offers in terms of control were investigated. A calibration was proposed, and its ability to be performed with minimal errors through visual feedback was examined. The type of hardware that could provide the best visual feedback, as well as the tracking and detection algorithms that could provide the highest accuracy, were explored. Both research and experimentation were carried out in order to address these statements. The Intel RealSense Depth Camera (D435i), a stereo solution, was implemented to provide vi-sual feedback. Four different trackers, CSRT, MOSSE, KCF and Medianflow, were implemented and compared to one another through testing. The tracker with the most satisfying results KCF was then chosen to perform a robot calibration. In the end, the testing showed that in the current implementation, the use of visual feedback provided unsatisfactory results. However, visual feedback could still prove feasible when implementing the necessary updates, such as a combination of IR and Aruco markers for higher accuracy and a combination of KCF and Kalman Filtering for better tracking results.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:50 technical science in general
Programme:Systems and Control MSc (60359)
Link to this item:https://purl.utwente.nl/essays/86070
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page