Automatic Optimal Design of a Visual-based Stiffness Sensor and real-time Colour-coded Stiffness Map for Minimally Invasive Procedures.

本研究的实施涉及两个系统的创建，一个放置在由机器人操纵器和基于视觉的感测机构组成的远程站点中，另一个放置在临床室中，其中操作者可以控制和重新规划机器人的运动，同时可视化所探索的解剖区域的实时刚度信息。UFactory Lite 6协作机器人已经被购买并与机器人操作系统（ROS）接口，ROS是一个强大的框架，允许实现机器人算法模拟并与真实的硬件轻松接口。半自主导航算法，其中操作员可以远程控制机器人的运动并重新规划其路径，已经使用ROS中的运动规划导航框架Moveit实现。虽然操作员可以重新规划机器人的运动并远程控制它，虚拟夹具将使用可视化实现-刚度信息，以控制力，并在接触过程中整合安全期货。控制和规划算法已初步测试的模拟，现在接口与真实的机器人进行仿真，提出了一种新的三维机器学习算法，打破了传统方法的局限性，实现了传感机构的优化设计。该方法仍处于开发阶段。

The implementation of this research is related to the creation of two systems, one placed in the remote site composed by a robot manipulator and a visual based sensing mechanism, and another placed on the clinical room in which an operator can control and replan the motion of the robot while visualizing real-time stiffness information of the explored anatomical area.A light weight six degree of freedom manipulator, the UFactory Lite6 collaborative robot, has been purchased and interfaced with the Robot Operating System (ROS), a powerful framework which allow implementation of algorithms-simulations for robotics and easy interface with real hardware.A semi-autonomous navigation algorithm, in which the operator can control the motion of the robot remotely and replan the its path, has been implemented using Moveit, a motion planning navigation framework, in ROS.Although the operator can replan the motion of the robot and control it remotely, virtual fixture will be implemented using the visual-stiffness information to control the force and integrate safety futures during the contact.The control and planning algorithm have been initially tested in simulations and are now interfaced with the real robot.A novel 3D machine learning algorithm, that shatters the constraints of conventional methodologies, is implemented to find the optimal design of the sensing mechanism. The method is still in a development phase.