|
|
|
| e-Pub |
Previous | 
Home
Section: New Results
Medical Robotics
Visual Servoing using Wavelet and Shearlet Transforms
Participants : Lesley-Ann Duflot, Alexandre Krupa.
In collaboration with Femto-ST lab in Besançon and the Research Group on Computational Data Analysis at Universitat Bremen, we developed a new generation of direct visual servoing methods in which the signal control inputs are the coefficients of a multiscale image representation. In particular, we considered the use of multiscale image representations that are based on discrete wavelet and shearlet transforms. We succeeded to derive an analytical formulation of the interaction matrix related to the wavelet and shearlet coefficients and experimentally demonstrated the performances of the proposed visual servoing approaches. We also considered this control framework in the design of a medical application which consists in automatically moving a biological sample carried by a parallel micro-robotic platform using Optical Coherence Tomography (OCT) as visual feedback. The objective is to automatically retrieve the region of the sample that corresponds to an initial optical biopsy for diagnosis purpose. First results obtained with a 3 DoF eye-to-hand visual servoing demonstrated the feasibility to use the wavelet coefficients of the OCT image as input of the control law.
3D Steering of Flexible Needle by Ultrasound Visual Servoing
Participants : Jason Chevrie, Marie Babel, Alexandre Krupa.
We pursued our work on 3D steering of a flexible needle using ultrasound visual servoing [11]. This year, in collaboration with the Surgical Robotics Laboratory of the University of Twente, we developed a method to control a 2 DoF needle insertion device attached to the end-effector of a 6-DoF robotic arm in order to automatically insert a flexible needle toward a spherical target embedded in a moving biological tissue (bovine liver). We proposed a method that uses both base manipulation control and tip-based control while compensating the tissue motion to avoid lateral tearing. The visual feedback provided by the ultrasound probe was used to track the target and an electromagnetic tracker attached inside the needle was used to locate its tip. In this study, the motion compensation of the moving tissue was performed by minimizing the interaction force measured at the base of the needle insertion device. In our approach we used the generic task control framework to fuse the needle targeting and motion compensation tasks into a single control law. First experimental ex-vivo results demonstrated the efficiency of the proposed control to reach a target in moving biological tissue.
Robotic Assistance for Ultrasound Elastography by Visual Servoing, Force Control and Teleoperation
Participants : Pedro Alfonso Patlan Rosales, Alexandre Krupa.
This work concerns the development of a robotic assistant system for quantitative ultrasound elastography. This imaging modality provides the elastic parameters of a tissue which are commonly related with a certain pathology. It is performed by applying continuous stress variation on the tissue in order to estimate a strain map. Usually, this stress variation is performed manually by the user through the manipulation of the ultrasound probe and it results therefore in an user-dependent quality of the strain map. To improve the ultrasound elastography imaging and provide quantitative measurement, we developed an assistant robotic palpation system that automatically moves a 2D ultrasound probe for optimizing ultrasound elastography [72]. This year we extended our previous robotic palpation system in order to perform 3D elastography and allow the user to teleoperate the probe orientation through a haptic device [56]. This extension is based on the use of a 3D ultrasound probe held by a 6 DoF robotic arm and the design of a new control law based on the task control framework that simultaneously performs three tasks: i) autonomous palpation by force control of the tissue required for the strain map estimation, ii) probe lateral alignment on a stiff target of interest for optimizing its visibility by visual servoing and iii) teleoperation of the probe orientation by the user for exploration purpose. Recently, we also proposed a solution that allows the estimation of the strain map of a moving tissue that is subject to physiological motion [57]. It is based on the combination of a non-rigid motion tracking of the tissue of interest in the ultrasound image and an automatic 6 DoF compensation of the perturbation motion by visual servoing using dense ultrasound information.
Haptic Guidance of a Biopsy Needle
Participants : Hadrien Gurnel, Alexandre Krupa.
We started a new study in collaboration with Maud Marchal (Inria Hybrid group) related to the assistance of manual needle steering for biopsies or therapy purposes (see Section 9.1.7). Instead of automatically inserting the needle by a robotic arm as we did in other works, our objective is to develop a solution that provides haptic cue feedback to the clinician that helps him during its manual gesture. The haptic cue feedback will be provided by a haptic device holding the needle. This year we developed a software tool that simulates and visualizes the interaction of a virtual needle with a deformable virtual organ. This organ is represented by a 3D mesh and a mass-spring-damper model was considered to simulate its deformation due to the needle insertion motion. The development of this software was based on our libraries UsTk and ViSP and the external library VTK (Visualization Toolkit). We also interfaced to this simulator our Haption Virtuose 6D haptic device to allow the user to teleoperate the virtual needle and to feel the force applied by the needle on the virtual tissue. This simulator will constitute an important tool for our future development of dynamic haptic guides before testing them in a real experimental setup.