Robotic Surgical Skull Cutting!

surgical roboticscontrolspHRI
Robotic Surgical Skull Cutting!
View on GitHub Live Demo

Note

Due to confidentiality agreement, my work cannot be expressed in detail until our paper has been accepted to the conference ~ Jan 2026. Hence the github link goes to the demo video

Overview

The goal of this project was to provide a reliable robotic solution to collaborative Craniosynostosis operations in fetal patients. Prior work can be found in the paper linked

The system is comprised of the custom tendon driven continuum/snake mechanism, connected to a Franka Research 3 serial robotic arm. Using the respective sensors and motors for the combined robotic system paired with a ROS middleware, it was possible to create a fully functioning User Interface for and low-level controls suite for both teleoperation thourgh mouse and keyboard, as well as collaborative movement through a force-guided controller.

Technical Details

  • System Modelling: To develop a useful and realistic force guided controller a 2nd order model was used to create virtual mass and damping to adjust the feel and sensitivity of the robot’s motion to the user force input.

  • Sensor Interfacing: Using different sensors in the control loop for accurate results involved post-processing such as filtering and smoothing, as well as applying wrench transformations to backwards solve input force based on the measurement sensor positioning.

  • Motor Drivers/Robot Interfaces: To properly send commands from the User Interface for either Teleoperation or Shared control, proper interfaces and layers were designed so make the user interface agnostic to operation mode (Sim/Real/Dig. Twin) while maintaining a real-time system with no delays and minimal overhead on process threads for different functions

Results

The results of the findings showed promising and accurate results with a generally low and acceptable error, with repeatable and predictable solutions. Further testing and results regarding the system will be found in the paper submission - once public.

Skills/Technologies Used

  • Python
  • C++
  • ROS
  • libfranka
  • Jax
  • Scipy
  • Numpy