An Improved RRT* Algorithm Combining Motion Constraint and Artificial Potential Field for Robot-Assisted Flexible Needle Insertion in 3D Environment

Abstract

In the field of minimally invasive medical care, compared With rigid needles, the use of flexible needles allows doctors to avoid blood vessels and organs more flexibly during surgery. Due to the non-holonomic constraint of the flexible needle and the tissue deformation caused by the needle tip during the insertion, the accessibility and safety of the path need to be considered when planning. In this paper, we propose an adaptable algorithm based on RRT* algorithm to compute a path for flexible needle to reach the lesion in 3D environment. The improved RRT* (Rapidly-Exploring Random Trees*) algorithm based on the motion constraint of flexible needle makes the path smoother and shorter. The path cost used in the algorithm take the potential field of the obstacle into account to avoid the influence of the local movement of the tissue during the needle puncture process. Finally, the algorithm is conducted by simulations, and the results show that the improved RRT* algorithm can generate a smooth and safe path that conforms to the motion constraint of the flexible needle.