Establishment of Estimation method for Large-Sized Object Inertia Parameters By Using Human-Skill-Like Manipulation

在机器人稳定、精确地操纵物体的许多情况下，需要知道物体的质量和质心。对于物体的重量或形状超过机器人手的抓取能力的情况，本研究于2001年提出了一种技术，该技术可以估计具有曲面和基平面的无抓取未知物体的质量和质心。一条叫做Toward-C. M的线路。向量及其对应的直线，Passing-C.M.定义了包含质心的直线。用于估计通过C.M.线，提示操作的机器人手指，提示对象缓慢和重复的平行运动与垂直的操作平面，提出。提出了一种利用指尖操作测量得到的指尖位置和力信息来估计传递-C. M.线描述。然后给出了一种通过估计多个不同方向的交叉点来估计物体质量和质心的算法。线最后，对所提出的方法进行实验验证，其结果进行了概述。在2002年，本研究提出了一种方法，估计物体的惯性参数在物体推操作的机器人。在推进操作中，由于物体的重力被其支撑表面抵消，所以物体上的重力变化的信息不能被机器人感知。本研究考虑推动一个物体，改变其加速度，然后通过机器人手指感知物体上的力的变化，这是随着物体的加速度变化，。利用物体的力和位置信息，提出了物体质量、质心和惯性矩的估计算法。并对所提出的方法进行了实验验证，其结果进行了概述。

In many cases of manipulating an object stably and accurately by robot, it is required to known the mass and center of mass of the object. For the case when weight or shape of an object is over the grasp capacity of a robot hand, in 2001, this research proposes a technique that can estimate the mass and center of mass of a graspless unknown object, which has curved surfaces and a base plane. A line called Toward-C.M. Vector and its corresponding line, Passing-C.M. Line, which contains the center of mass, are defined. For estimating the passing-C.M. line, Tip Operation by robot finger, which tips the object slowly and repeatedly in a parallel motion with a vertical operation plane, is proposed. Using the fingertip position and force information measured from tip operations, an algorithm to estimate the passing-C.M. line are described. Then an algorithm to estimate the mass and center of mass of the object is given by estimating the intersect point of several orientation-different passing-C.M. lines. Lastly, experimental verification on the proposed approach is performed and its results are outlined.In 2002, this research proposes a method for estimating the object inertia parameters in object pushing operations by robot. On pushing operation, the information of the gravity change on an object cannot be sensed by robot since the object gravity is cancelled by its supporting surface. This research considers pushing an object with changing its acceleration, then sensing the force variations on the object, which are changing with the object acceleration, by robot fingers. With the force and position information, algorithms for estimating the mass, center of mass and inertia moment of the object are addressed. And experimental verification on the proposed approach is performed and its results are outlined.