Optimization-based performance maximization and trajectory planning for robot manipulators최적설계를 기반으로 하는 로봇 매니퓰레이터의 성능 극대화 및 경로 계획

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 124
  • Download : 0
In this study, to maximize the performance of a robot manipulator optimization-based method are proposed. The concept of allowable velocity and force is proposed to precisely evaluate the maximum directional kinematic capability of a redundant manipulator. For a general redundant manipulator, an optimization problem is formulated to determine the maximum achievable velocity and force projected along the base direction at any target position in the workspace. This provides quantitative information on allowable (i.e., maximum directional) velocity and force to be precisely visualized in 2D and 3D complicated shapes, which conventional manipulability ellipsoid cannot provide. As application examples, allowable velocity and force are evaluated for the distributed actuation mechanism (DAM)-based three-link planar manipulator, the 3RRR planar parallel manipulator, and the UR5 robot (a spatial manipulator with 6 degrees of freedom). The simulation and experimental results validate that the proposed method can precisely determine allowable velocity and force, thereby contributing to planning the optimal operation for a given task.Simultaneous maximization of velocity and force of a robot manipulator is determined based on multi-objective optimization framework. To show the effectiveness of the proposed method, we applied it to a DAM-based three-link planar manipulator, the 3RRR planar parallel manipulator, and the UR5 robot. In particular, a novel concept of bioinspired variable gearing of the DAM was demonstrated to show the effectiveness of the proposed method. This study optimizes the performances (i.e., both velocity and force) of the DAM-based on the novel concept of continuously variable gearing, which is inspired by muscle movement. To quantify continuously variable gearing in the DAM, the structural gear ratio (defined as joint speed/motor speed) is mathematically derived in terms of the slider position and the joint angle. Then, for a DAM-based three-revolute joint manipulator, a multi-objective optimization problem is formulated to determine the maximum end-effector velocity according to varying payloads. An optimization framework consisting of the analysis and optimization modules is constructed to verify the proposed concept with a comparison of an equivalent joint actuation mechanism (JAM)-based three-revolute joint manipulator. The numerical results demonstrate that the bioinspired variable gearing of the DAM allows for a significant enhancement of end-effector velocity and force, depending on a given task.In addition, trajectory planning based on the proposed maximum performance evaluation method was conducted for a manipulator based on the DAM and the equivalent JAM. The overall path was set based on the allowable velocity and force polygons, and the specific task was set based on the result of multi-objective optimization. Then, trajectory planning was determined through linear interpolation based on the optimization results for each target positions, and numerical analysis was performed for the tracking control. As a result, it was shown that effective trajectory planning was performed through the proposed maximum performance evaluation method.
Advisors
Jang, In Gwunresearcher장인권researcher
Description
한국과학기술원 :조천식녹색교통대학원,
Publisher
한국과학기술원
Issue Date
2021
Identifier
325007
Language
eng
Description

학위논문(박사) - 한국과학기술원 : 조천식녹색교통대학원, 2021.2,[iv, 58 p. :]

Keywords

Optimization▼aAllowable velocity and force▼aMaximum performance evaluation▼aBioinspired variable gearing▼aTrajectory planning; 최적화▼a허용 가능 속도 및 힘▼a최대 성능 평가▼a생체모방 가변 기어링▼a경로 계획

URI
http://hdl.handle.net/10203/309301
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1006567&flag=dissertation
Appears in Collection
GT-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0