CRANE: a 10 Degree-of-Freedom, Tele-surgical System for Dexterous Manipulation within Imaging Bores

Dimitri A. Schreiber, Zhaowei Yu, Hanpeng Jiang, Alexander M. Norbash, Michael C. YipICRA 2021 · OpenReview

TL;DR

This paper introduces CRANE, a novel 10-DoF robotic system designed for highly dexterous and accurate needle insertion within CT scanner bores. By addressing limitations in existing robotic platforms, CRANE aims to improve the safety and efficiency of minimally invasive procedures like biopsies and ablations.

What Problem Does This Paper Address?

Minimally invasive procedures, particularly those guided by CT imaging for targeting sub-dermal lesions in areas like the lungs and liver, are crucial for diagnosis and treatment. However, current methods often involve physicians manually inserting needles, requiring them to repeatedly withdraw from the scanner to get updated imaging. This process is time-consuming, prone to inaccuracies due to the lack of real-time feedback and tool tracking, and can lead to increased radiation exposure and patient discomfort. Existing robotic solutions often suffer from limited dexterity, workspace, or accuracy within the confined and artifact-prone environment of a CT scanner.

Methods

The CRANE system is a highly dexterous, low-profile, 8+2 Degree-of-Freedom (DoF) robotic arm designed for CT-guided percutaneous needle biopsy. Its core innovation lies in its in-bore mechanical design, which prioritizes minimizing volume within the scanner. It utilizes a remote actuation system with 2N cable drives and lightweight materials like carbon fiber and ceramic bearings to reduce inertia and prevent imaging artifacts.

A key feature is the novel needle interface, employing a disposable, removable, and sterilizable mechanism with Shape Memory Alloy (SMA) wire actuators. This "inch-worm" like mechanism allows for precise needle manipulation and grasping of various needle sizes.

For control, CRANE employs a multi-level control scheme. This includes a low-level system with a DE0-NANO-SoC and FPGA for real-time PID control of joint motors, and a high-level ROS-based system running on a desktop PC for user interface and end-effector control. A complementary filter is used to fuse motor velocity and magnetic joint encoder data for accurate joint state estimation, mitigating noise and coupling matrix errors. End-effector pose errors are calculated and used to update joint setpoints via a PD control law, incorporating Jacobian pseudoinverses for inverse kinematics.

Datasets

Not discussed in the paper. The evaluation relies on trajectory tracking tests and a CT scanner biopsy experiment using a custom lung phantom.

Key Insights and Ideas

  • Dexterity within the Bore is Paramount: The paper emphasizes that achieving high dexterity inside the CT scanner is critical for navigating complex anatomy and avoiding collisions, rather than just relying on external robotic arms.
  • SMA Actuation for Needle Grasping: The use of SMA wires for a disposable needle clutch is an elegant solution for secure, yet easily releasable, needle manipulation, addressing sterilization and adaptability concerns.
  • Multi-Level Control for Accuracy: Combining low-level joint control with high-level end-effector feedback and complementary filtering is essential for overcoming the inherent challenges of robotic systems in noisy, artifact-prone environments like CT scanners.
  • Minimizing Artifacts is Non-Negotiable: The careful selection of materials (plastics, composites, ceramics) and avoidance of high-density components within the bore is crucial for maintaining image quality during procedures.

Key Results

  • Trajectory Tracking Accuracy: The system achieved a mean tracking error of < 0.2mm for position and < 0.71° for orientation during trajectory tracking evaluations.
  • Needle Clutch Performance: Activated slipping forces were measured at 18N and 20N, with deactivated slipping forces at 1N and 2.25N. The clutch demonstrated rapid on/off times, with a 2.5-second rise time and 10.1-second fall time for temperature control.
  • CT Biopsy Experiment: The system successfully performed needle insertions into a custom lung phantom, with insertion vectors having less than 1.2° error from nominal. Minimal shadowing and artifacts were observed in the CT images.
  • Hardware Performance: Hardware stiffness was measured at > 100N/mm, and backlash was < 0.1mm, significantly exceeding clinical requirements.

Limitations and Future Work

  • Limitations:
    • The paper acknowledges that FEA analysis showed a 10.3mm suggested tip deflection under a 10N force, motivating the need for direct end-effector sensing and feedback control.
  • Open-loop control without feedback resulted in significantly increased position and orientation errors, highlighting the necessity of closed-loop control.
  • The current system's interference with the scanner and magnetic tracker was minimal, but further investigation into potential interference is always warranted.
  • Future Work:
    • Developing collision-free path planning algorithms specifically for CT scanner environments.
  • Exploring metrics for ranking inverse kinematics solutions to optimize target needle-insertion vectors.
  • Utilizing null-space control to better leverage the robot's dexterity and avoid joint limits.

Summary updated

Abstract

Physicians perform minimally invasive percutaneous procedures under Computed Tomography (CT) image guidance both for the diagnosis and treatment of numerous diseases. For these procedures performed within Computed Tomography Scanners, robots can enable physicians to more accurately target sub-dermal lesions while increasing safety. However, existing robots for this application have limited dexterity, workspace, or accuracy. This paper describes the design, manufacture, and performance of a highly dexterous, low-profile, 8+2 Degree-ofFreedom (DoF) robotic arm for CT guided percutaneous needle biopsy. In this article, we propose CRANE: CT Robot and Needle Emplacer. The design focuses on system dexterity with high accuracy: extending physicians' ability to manipulate and insert needles within the scanner bore while providing the high accuracy possible with a robot. We also propose and validate a system architecture and control scheme for low profile and highly accurate image-guided robotics, that meets the clinical requirements for target accuracy during an in-situ evaluation. The accuracy is additionally evaluated through a trajectory tracking evaluation resulting in <0.2mm and <0.71degree tracking error. Finally, we present a novel needle driving and grasping mechanism with controlling electronics that provides simple manufacturing, sterilization, and adaptability to accommodate different sizes and types of needles.
BibTeX
@inproceedings{schreiber2021crane,
  author = {Dimitri A. Schreiber and Zhaowei Yu and Hanpeng Jiang and Alexander M. Norbash and Michael C. Yip},
  title = {{CRANE: a 10 Degree-of-Freedom, Tele-surgical System for Dexterous Manipulation within Imaging Bores}},
  booktitle = {IEEE International Conference on Robotics and Automation},
  year = {2021},
  url = {https://openreview.net/forum?id=4P9sSTNcX9},
}
Figure