Machine Learning meets Continuum Robotics

Continuum robots exhibit a curvilinear morphology which is compliant. Common robot modelling techniques to describe the relationship between actuator input to the position and orientation in task space, and vice versa, are thus not applicable. Establishing a physics-based model capturing the kinetostatic behavior of continuum robots is challenging as elastic interactions result in a highly non-linear problem. The Goldstandard physics-based model using the Cosserat theory of elastic rods achieves reasonable approximations with 1.5 − 5% with respect to the robot’s length, if the model parameters are well tuned and the system is well-calibrated.

Case Study: Concentric Tube Continuum Robots

Concentric tube continuum robots (CTCR) are composed of multiple concentric, pre-curved super-elastic tubes that are rotated and translated relative to one another. Due to the elastic interactions between the tubes, highly non-linear behavior characterizes the mapping between task space and joint space of a CTCR. To reflect this interaction, model-based approaches utilize static models based on the theory of Cosserat rods and are formulated as a boundary value problem which is solved numerically. In favour of computation time, some physical phenomena, e.g. friction or tube tolerances, are commonly neglected in these Goldstandard physics-based modeling approaches. This model-­based approach sufficiently represents the underlying mechanics, but has limitations in terms of accuracy as well as efficiency. An inversion of the model can only be realised numerically by iteratively solving the forward model at high computational expense. High computational expense is also characterising state of the art methods in computational design optimization and motion planning, while not exploring the complete parameter space and thus leading to suboptimal results. Learning-based models of concentric tube continuum robots have outperformed the Goldstandard Cosserat rod model with approximation errors below 1%!

Model-based vs. Data-based Methods

The aim of this research programme is to leverage learning-based approaches for modelling, computational design, and motion planning for continuum robots. The long term vision is to enable technology transfer of these techniques to real-­world applications of these robot, such as minimally invasive surgery, by focusing on the knowledge gaps for step change research. In the effort of achieving this, this research programme is planned around four objectives: - Define Data Representations - Investigate Deep Learning for Robot Modelling - Enable Task-­Optimal Robot Designs by Reinforcement Learning - Explore Learning­-based Motion Planning Techniques


  • NSERC Discovery Grant (April 1, 2019 – March 31, 2025)
  • NSERC Discovery Accelerator Supplements
  • NSERC Discovery Launch Supplements (2019 - 2020)


Publication Image
Design of Patterns in Tubular Robots Using DNN-Metaheuristics Optimization

Soyeon Park, Jongwoo Kim, Jieun Park, Jessica Burgner-Kahrs, Gunwoo Noh

International Journal of Mechanical Sciences, 251:108352, 2023.

Publication Image
A Dataset and Benchmark for Learning the Kinematics of Concentric Tube Continuum Robots

Reinhard M. Grassmann, Ryan Z. Chen, Nan Liang, Jessica Burgner-Kahrs

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022.

Publication Image
Learning-based Inverse Kinematics from Shape as Input for Concentric Tube Continuum Robots

Nan Liang, Reinhard M Grassmann, Jessica Burgner-Kahrs

IEEE International Conference on Robotics and Automation (ICRA), 2021.

Publication Image
Estimating Tip Contact Forces for Concentric Tube Continuum Robots based on Backbone Deflection

Heiko Donat, Sven Lilge, Jessica Burgner-Kahrs, Jochen J Steil

IEEE Transactions on Medical Robotics and Bionics , 2 (4), pp. 619-630, 2020.

Publication Image
On the Merits of Joint Space and Orientation Representations in Learning the Forward Kinematics in SE(3)

Reinhard Grassmann, Jessica Burgner-Kahrs

Robotics - Science and Systems Conference, 10 pages, 2019.

Publication Image
Towards Learning Force Sensing for a Concentric Tube Continuum Robot

Heiko Donat, Sven Lilge, Jochen Steil, Jessica Burgner-Kahrs

Workshop on Open Challenges and State-of-the-Art in Control System Design and Technology Development for Surgical Robotic Systems, IEEE International Conference on Robotics and Automation, 2019.

Publication Image
Learning the Forward and Inverse Kinematics of a 6-DOF Concentric-Tube Continuum Robot in SE(3)

Reinhard Grassmann, Vincent Modes, Jessica Burgner-Kahrs

IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 5125-5132, 2018.