Smooth on-line path planning for needle steering with non-linear constraints

摘要

Percutaneous intervention is a commonly used surgical procedure for many diagnostic and therapeutic operations. Target motion in soft tissue during an intervention caused by tissue deformation is a common problem, along with needle displacement. In this work, we present a deformation planner that generates continuous curvature paths with a bounded curvature derivative that can be used on-line to reach a moving target. This planner is computationally inexpensive and can be used for any robotic system, which has finite angular velocity, to reach a mobile target. The deformation planner, is integrated into a needle steering system using a novel, biologically inspired needle, STING, to track a simulated moving target. In-vitro results in gelatin demonstrate accurate 2D tracking of a moving target (mean 0.27 mm end positional error and 0.80° approach angle error) over 3 target movement rates.