Hybrid control of needle insertion using visual servoing

Contact: Jason Chevrie
Last update: July 2017

Context

This project aims at providing robotic assistance for needle insertion procedures, in which typically long and thin needles need to be steered in soft tissues. Accurate steering of the needle toward a target is difficult to perform because of the complex interaction of the flexible needle with the tissues. Additionally such needles usually have a bevel at their tip to facilitate tissue cutting. This modifies the trajectory of the needle tip, even when the needle is apparently inserted straight into the tissues. Using a robotic system to control the needle insertion would ease the task for the surgeon.

Method overview

We use an online 3D model of flexible needle in interaction with soft tissues to model the current state of the insertion. Using this model, we design a control law to compute the motion to apply to the needle base in order to steer the tip toward a target region. Close-loop control is performed by tracking the position of the needle in visual feedback. Two different controls are used:

• Lateral base manipulation: the needle base is moved laterally to create needle bending and tissue deformation to modify the shape of the needle and align the tip axis with the target. This control is primarily used when the needle is misaligned with the target and could not be reached accurately if the needle was inserted straight in the tissues.
• Duty cycling: a combination of insertion with rotation around the needle axis and insertion without rotation is used to control the deviation of the needle tip trajectory. This control is used when the target is almost aligned with the needle and fine control of the needle tip deviation is sufficient to reach the target.

Tissue motion also induce needle bending outside the tissue which can modify the tip trajectory. Visual tracking of the needle shaft is used to update online the position of the tissue in the interaction model. The model can then be used to provide a prediction of the needle trajectory and a more accurate control of the needle trajectory.

Experimental results

References

1. J. Chevrie, A. Krupa, M. Babel. Needle steering fusing direct base manipulation and tip-based control. In IEEE Int. Conf. on Robotics and Automation, ICRA'16, Pages 4450-4455, Stockholm, Sweden, May 2016.
2. J. Chevrie, A. Krupa, M. Babel. Online prediction of needle shape deformation in moving soft tissues from visual feedback. In IEEE/RSJ Int. Conf. on Intelligent Robots and Systems, IROS'16, Pages 2375-2380, Daejeon, Korea, October 2016