Purpose

Continuum manipulators are often used in complex and narrow space in recent years because of their flexibility and safety. Vision is considered to be one of the most direct methods to obtain its spatial shape. However, with the improvement of the cooperation requirements of multiple continuum manipulators and the increase of space limitation, it is impossible to obtain the complete spatial shape information of multiple continuum manipulators only by several cameras.

Design/methodology/approach

This paper proposes a fusion method using inertial navigation sensors and cameras to reconstruct the shape of continuum manipulators in the whole workspace. The camera is used to obtain the position information, and the inertial navigation sensor is used to obtain the attitude information. Based on the above two information, the shape of the continuum manipulator is reconstructed by fitting Bézier curve.

Findings

The experiment result of single continuum manipulator shows that the cubic Bézier curves is applicable to curve fitting of variable curvature, the maximum fitting error is about 2 mm. Meanwhile, the experiment result shows that this method is not affected by obstacles and can still reconstruct the shape of the continuum manipulators in 3-D space by detecting the position and attitude information of the end.

Originality/value

According to the authors’ knowledge, this is the first study on spatial shape reconstruction of multiple continuum manipulators and the first study to introduce inertial navigation sensors and cameras into the field of shape reconstruction of multiple continuum manipulators in narrow space. This method is suitable for shape reconstruction of manipulator with variable curvature continuum manipulator. When the vision of multiple continuum manipulators is blocked by obstacles, the spatial shape can still be reconstructed only by exposing the end point. The structure is simple, but it has certain accuracy within a certain range.

中文翻译：

---

使用惯性导航传感器和相机的多连续体机械手形状重建

目的

连续体机械手因其灵活性和安全性，近年来常用于​​复杂、狭窄的空间。视觉被认为是获得其空间形状最直接的方法之一。然而，随着多个连续体机械臂协作要求的提高和空间限制的增加，仅通过多个摄像头已经无法获得多个连续体机械臂的完整空间形状信息。

设计/方法/方法

本文提出了一种使用惯性导航传感器和相机的融合方法来重建整个工作空间中的连续体机械手的形状。相机用于获取位置信息，惯性导航传感器用于获取姿态信息。基于以上两个信息，通过拟合贝塞尔曲线重建连续体机械手的形状。

发现

单个连续体机械手的实验结果表明三次贝塞尔曲线适用于变曲率的曲线拟合，最大拟合误差约为2 mm。同时，实验结果表明，该方法不受障碍物的影响，仍然可以通过检测末端的位置和姿态信息来重建三维空间中连续体机械臂的形状。

原创性/价值

据作者所知，这是首次对多连续体机械臂空间形状重建的研究，也是首次将惯性导航传感器和相机引入狭小空间多连续体机械臂形状重建领域。该方法适用于变曲率连续体机械臂的形状重建。当多个连续体机械臂的视野被障碍物挡住时，仍然可以通过暴露端点来重建空间形状。结构简单，但在一定范围内具有一定的精度。