IEEE ICRA 2012 Needle Steering Workshop

Control and Estimation for Steerable Needles

John P. Swensen1, MingDe Lin2, Allison M. Okamura3, Noah J. Cowan4

1 Yale University

2 Philips Healthcare

3 Stanford University

4 Johns Hopkins University

Abstract

The problem of estimation and control of steerable needles is a critical component of moving robotic interventions from the testbed to the clinic. The ability to accurately determine placement, correct for errors due to anatomy and disturbance, and assure safety and repeatability are necessary to show effectiveness of robotic interventions. In this talk, we first present an overview of the key challenges associated with estimating and controlling steerable needles. The problem of estimation depends heavily on the sensing modalities available for each particular medical intervention. In the second part we discuss a number of techniques we have applied to solve the problems of estimation and control separately and discuss how to combine them in the future.  We present results from our closed-loop experiments under stereo camera feedback. In addition, we demonstrate successful closed-loop estimation and control methods of tip-steerable needles using C-arm fluoroscopy. We conclude that our methods work well both with stereo vision and medical imaging feedback.

 



Convergence plots for a variety of trials conducted under different imaging modalities where the goal is to control the tip of the needle to a plane. Here, the distance from the plane is plotted: the solid line indicates the average error across multiple trials and the shaded region being the point-wise standard deviation. (A) Experiments conducted on our bench-top testbed with triangulate tip location from stereo cameras giving the distance to the plane. (B) Experiments conducted with the Allura Xper FD20 (Philips Healthcare, Best, The Netherlands) where the scaled distance from the centerline of the imager is the distance to the plane. As can be seen, the results are comparable despite dramatically different imaging modalities. 

Related Publications

Links

http://www.lcsr.jhu.edu/Ns/Overview