Visual/Inertial Detection of Trocar and Remote Centre of Motion in Vitreoretinal Surgery

点击翻译按钮获取中文摘要

This PhD project pertains to improving safety in keyhole micro-surgery. Keyhole micro-surgery takes place under strict precision requirements, and, in many cases, requires dexterity that is at the limits of human capabilities. One such example is vitreoretinal surgery, where extremely small structures on the order of the human hair must be manipulated. Upcoming vitreoretinal surgical interventions would require precision even below that, so that regenerative therapies can be delivered to specific retinal layers, which have a thickness on the order of 10-20um.In vitreoretinal surgery, the clinician manipulates an instrument inserted through a trocar/hole on the eye that acts as the fulcrum point. Therefore, a remote centre of motion should be formed at this fulcrum point, as otherwise there is the risk of trauma as the tool pulls on the tissue rather than pivoting. For example, there is the danger that the eye's tissue will be pulled apart, causing, for example, astigmatism or lens displacement. When such surgeries are manually performed, the clinician uses haptic feedback in order to understand that excessive forces are applied. As vitreoretinal surgery slowly considers the use of robotics, it is important to develop mechanisms to understand whether the fulcrum point and remote centre of motion are identical. Knowledge of this improves safety but also allows the bimanual manipulation of suspended organs such as the eye. While there is research on developing force sensors that understand the amount of applied force, we are interested in developing a visual/inertial approach to tackle this problem. This would require the minimum customisation of existing instruments, therefore enabling their use from upcoming robotic assistants.Specifically, the key questions in this PhD project are:1) Can existing vitreoretinal surgical instruments we retrofitted in an acceptable fashion with miniature cameras and inertial sensors to support the development of such an approach?2) Can Simultaneous Localisation and Mapping (SLAM) be used in conjunction with miniature cameras and inertial measurements to localise the trocar and calculate the remote centre of motion?3) Can the information about the remote centre of motion and trocar be provided to a robotic assistant to improve safety?