Path Planning with Modified RRT* Algorithm for Lung Biopsy

Abstract

– Path planning plays a central role in robot-assisted percutaneous insertion. The main challenge of path planning exists in the motion constraints inherited from the geometry and mechanics of the needle, and the complex anatomic environment of human body. In nonholonomic planning, the classic Rapidly-Exploring Random Trees (RRT) algorithm may fail to provide a continuous and obstacle-avoidable path. To find a feasible path and minimize the damage on soft tissues based on a newly-introduced curvature-controllable steerable needle, we propose a method that utilizes RRT* and quadratic Bezier curve smoothing technique. RRT* with Bezier Curve Smoothing can generate a path composed of smooth piecewise planar curves with continuous connections. Comparisons are employed to show that our method generates shorter and less torturous paths with a higher success rate.