Task-oriented design of concentric tube robots using mechanics-based models

Abstract

We introduce a method for task-oriented design of concentric tube robots, which are tentacle-like robots with the potential to enable new minimally invasive surgical procedures. Our objective is to create a robot design on a patient-specific and surgery-specific basis to enable the robot to reach multiple clinically relevant sites while avoiding anatomical obstacles. Our method uses a mechanically accurate model of concentric tube robot kinematics that considers a robot's time-varying shape throughout the performance of a task. Our method combines a search over a robot's design space with sampling-based motion planning over its configuration space to compute a design under which the robot can feasibly perform a specified task without damaging surrounding tissues. To accelerate the algorithm, we leverage design coherence, the observation that collision-free configuration spaces of robots of similar designs are similar. If a solution exists, our method is guaranteed, as time is allowed to increase, to find a design and corresponding feasible motion plan. We provide examples illustrating the importance of using mechanically accurate models during design and motion planning and demonstrating our method's effectiveness in a medically motivated simulated scenario involving navigation through the lung.