Efficient 3D control for needle steering using duty-cycled rotation

Abstract

Bevel-tipped flexible needles can be steered to reach clinical targets along curvilinear paths in 3D while avoiding obstacles. Steering by duty-cycled rotation increases the versatility of this approach by providing proportional control of trajectory curvature. This paper presents computationally efficient techniques for path planning and path-following control for this application, using a 3D simulated brain environment. Path planning algorithms for this class of steerable needles have been developed using Rapidly-exploring Random Trees (RRTs). This paper expands on these methods, using quaternions for representation of rotation, and enhancing computational efficiency through use of interpolation, and by relaxing the entry constraint. For path-following, a look-ahead proportional controller for position and orientation is presented. Simulations in a 3D brain-like environment demonstrate the performance of the proposed planner and path-following controller. The look-ahead is seen to improve path-following performance.