Multi-Target Planning of Steerable Tubular Robots for Medical Procedures

Advances in robotics have enabled the creation of concentric tube robots that resemble a flexible bevel-tip needle. These robotic devices can be steered around obstacles, allowing them to complete complex medical procedures, such as biopsies and brachytherapy, with minimal invasiveness. These procedures often require the robotic needle to traverse living tissue to reach a set of designated areas while avoiding sensitive regions. Furthermore, the robot dynamics limit how much the needle can bend at any given point. As such, the optimal path for the device will traverse all target locations while minimizing tissue trauma, avoiding sensitive regions, and satisfying the robot dynamics. We show that optimal control theory can be applied to resolve the challenge of finding such a path. We pose path planning as a constrained optimization problem. Our cost functional consists of two components: a cost associated with the path length, and a cost associated with the distance of the needle from a set of target locations. The path planner finds a path that satisfies the robot dynamics and minimizes the cost functional. We analyze the efficacy of the path planner by simulating path generation in several 2D environments that are loaded with target locations and sensitive regions. The generated path is compared to the result obtained by exhaustively testing a family of curves to find the best path for a particular environment. Our path planner is currently capable of determining the optimal path to a single target location and preliminary results are presented for multi-target scenarios.