Periodic Reporting for period 4 - PIONEER (Peri-Ocularly Navigated Exteroceptive Snake Robot for Novel Retinal Interventions)
Reporting period: 2021-07-01 to 2023-03-31
This project pertained the development of a flexible robot that navigates around the eye, within the orbit, to reach the back of the eye where it can deliver therapies to the optic nerve, eye vasculature, or retina. The current mechanism to access this part of the eye is either through the nose, requiring an invasive neurosurgical procedure, through catheterisation when the vessels need to be accessed, or the dislocation of the eye to enable access to its posterior.
By developing small robots that can minimally invasively reach the posterior of the eye with high dexterity, the project set the foundations for a new era of orbital interventions. The project delivered three such prototypes, each with three dexterous arms (one holding an endoscopic camera for visualisation, and two holding tools). In collaboration with leading ophthalmic surgeons, one of the prototypes has been co-designed to comprise a bespoke collimator that keeps the separation of the arms constant. This enabled easier reach to the optic nerve, where currently risky procedures such as Optic Nerve Sheath Fenestration can be carried out. Such procedures are currently challenging even for the most skilled of surgeons, making the development of dexterous robots a worthwhile pursuit.
The overall objectives of the project, apart from the design of the system, were to create a interventional image analysis algorithm that assists with guiding the robot to the location of interest, as well as a force sensing sleeve that covers the robot surface and understands forces applied to the anatomy with the goal to minimise them.
With regards to the objectives, the project delivered the following (non exhaustive) list of outputs:
- Introduction and modelling of the concept of eccentric tube robots, wherein multiple flexible arms are housed within a single lumen.
- Introduction and modelling of hybrid continuum robots, wherein a flexible backbone houses also flexible eccentrically arranged arms.
- Introduction and design optimisation of a multi-arm robot for optic nerve sheath fenestration.
- Reduced order models for the above and other flexible robots.
- Interventional image analysis algorithms with regards to optical flow, segmentation, and image quality enhancement.
- Visual servo control elements for surgical robots.
- Intelligent robot controllers that optimise force application by changing their shape.
By developing small robots that can minimally invasively reach the posterior of the eye with high dexterity, the project set the foundations for a new era of orbital interventions. The project delivered three such prototypes, each with three dexterous arms (one holding an endoscopic camera for visualisation, and two holding tools). In collaboration with leading ophthalmic surgeons, one of the prototypes has been co-designed to comprise a bespoke collimator that keeps the separation of the arms constant. This enabled easier reach to the optic nerve, where currently risky procedures such as Optic Nerve Sheath Fenestration can be carried out. Such procedures are currently challenging even for the most skilled of surgeons, making the development of dexterous robots a worthwhile pursuit.
The overall objectives of the project, apart from the design of the system, were to create a interventional image analysis algorithm that assists with guiding the robot to the location of interest, as well as a force sensing sleeve that covers the robot surface and understands forces applied to the anatomy with the goal to minimise them.
With regards to the objectives, the project delivered the following (non exhaustive) list of outputs:
- Introduction and modelling of the concept of eccentric tube robots, wherein multiple flexible arms are housed within a single lumen.
- Introduction and modelling of hybrid continuum robots, wherein a flexible backbone houses also flexible eccentrically arranged arms.
- Introduction and design optimisation of a multi-arm robot for optic nerve sheath fenestration.
- Reduced order models for the above and other flexible robots.
- Interventional image analysis algorithms with regards to optical flow, segmentation, and image quality enhancement.
- Visual servo control elements for surgical robots.
- Intelligent robot controllers that optimise force application by changing their shape.
With regards to WP1, computational robot design: Work has been conducted to have computationally efficient algorithms that evaluate the dexterity of continuum robots through a novel voxelisation approach. These algorithms are then used to inform the design of the robots given anatomical constraints. We are using those algorithms to design the robotic system.
WIth regards to WP2, robot engineering: Two robot prototypes have been computationally designed, engineering, and evaluated. One of them is a fully flexible peri-ocularly navigated robot, while the other is a hybrid approach that borrows elements from current surgical practice. Both comprise three arms: one carries a tiny endoscopic camera while two carry tools for the intervention.
With regards to WP3, robot control: We developed new robot manipulability metrics and performed hybrid force/position control of the robot tip.
With regards to WP4, image-based navigation: We are working on single-camera SLAM algorithms, but have no concrete results yet.
WIth regards to WP5, force sensing sleeve: We have developed a first force sensing sleeve prototype that can cover the exterior surface of a flexible robot.
WIth regards to WP2, robot engineering: Two robot prototypes have been computationally designed, engineering, and evaluated. One of them is a fully flexible peri-ocularly navigated robot, while the other is a hybrid approach that borrows elements from current surgical practice. Both comprise three arms: one carries a tiny endoscopic camera while two carry tools for the intervention.
With regards to WP3, robot control: We developed new robot manipulability metrics and performed hybrid force/position control of the robot tip.
With regards to WP4, image-based navigation: We are working on single-camera SLAM algorithms, but have no concrete results yet.
WIth regards to WP5, force sensing sleeve: We have developed a first force sensing sleeve prototype that can cover the exterior surface of a flexible robot.
The computational robot design, multi-arm robot proposal, control algorithms and force sensing sleeve have been published or are under consideration for publication. Therefore, they are now considered state-of-the-art. It is expected that until the end of the project, we will have:
- A robust fully actuated compact multi-arm robot that can be deployed on a phantom eye
- A set of control algorithms for multi-arm robots, enabling the exercise of high forces while maintaining high manipulability
- A simultaneous localisation and mapping (SLAM) algorithm enabling the semi-autonomous navigation of the robot as it is inserted in the eye orbit
- A force sensing sleeve that is flexible but precise enough to sense the forces applied by the robot to the environment in the orbit.
- A robust fully actuated compact multi-arm robot that can be deployed on a phantom eye
- A set of control algorithms for multi-arm robots, enabling the exercise of high forces while maintaining high manipulability
- A simultaneous localisation and mapping (SLAM) algorithm enabling the semi-autonomous navigation of the robot as it is inserted in the eye orbit
- A force sensing sleeve that is flexible but precise enough to sense the forces applied by the robot to the environment in the orbit.