• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2002년도 ICCAS
• /
• pp.92.1-92
• /
• 2002

$\textbullet$ A compact and robust wall-climbing robot, called MRWALLSPECT-II, is developed. $\textbullet$ The robot is a self-contained system for scanning external surfaces of gas or oil tank. $\textbullet$ The robot has advantages of reduced actuators, parts and easy control. $\textbullet$ The mechanism of the robot employs a closed link mechanism. $\textbullet$ Self-Contained, Wall-Climbing, MRWALLSPECT

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.2529-2534
• /
• 2003

A quadruped walking robot has a superior adaptablility as well as highly adaptable mobility in various environments. These special advantages are outstanding in the mobile robot group. In this work, we introduce the method for omni-directional gait and rotational gait which is the generalized control algorithm to perform any direction commands. In addition, to improve the stability of quadruped walking robot, we performed the optimization between walking angle and sequence of feet. The proposed ideas are applied to the actual design of MRWALLSPECT III(Multifunctional Robot for Wall inSpection version 3) that is designed to inspect of the large surface of industrial utilities. By implementing the proposed idea on the robot, it’s effectiveness is experimentally confirmed.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.639-644
• /
• 2003

An unstructured environment requires a robot to possess outstanding mobility and advanced control algorithms since there exist complicated configurations such as obstacle, uneven surface, etc. Especially, when a quadruped robot walks in these environments, obstacles in the walking route will obstruct the walking or may give rise to a serious trouble. In this paper, we introduce a strategy for the stable walking in unstructured environment. The proposed strategy consists of two control algorithms. One is a collision{free algorithm to avoid obstacles and the other is an algorithm to overcome any obstacle. These are based on the obstacle detection method and a shape reconstruction algorithm, Which algorithms are described in detail. In addition, the validity of these algorithms have been demonstrated through experiments using a quadruped walking robot called "MRWALLSPECT III(Multifunctional Robot for Wall inSPECTion version 3 )".

• Journal of Mechanical Science and Technology
• /
• 제18권4호
• /
• pp.573-581
• /
• 2004

A self-contained wall climbing robot, called MRWALLSPECT (Multi-functional Robot for WALL inSPECTion) II, is developed. It is designed for scanning external surfaces of gas or oil tanks with small curvature in order to find defects. The robot contains all the components for navigation in itself without any external tether cable. Although it takes the basic structure of the sliding body mechanism, the robot has its original characteristic features in the kinematic design with closed link mechanism, which is enabled by adopting a simple and robust gait pattern mimicking a biological system. By employing the proposed link mechanism, the number of actuators is reduced and high force-to-weight ratio is achieved. This paper describes its mechanism design and the overall features including hardware and software components. Also, the preliminary results of experiments are given for evaluating its performances.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2005년도 ICCAS
• /
• pp.314-319
• /
• 2005

When a robot navigates in the real environment, it frequently meets various environments that can be expressed by simple geometrical shapes such as fiat floor, uneven floor, floor with obstacles, slopes, concave or convex corners, etc. Among them, the convex corner composed of two plain surfaces is the most difficult one for the robot to negotiate. In this paper, we propose a gait planning algorithm to help the robot overcome the convex environment. The trajectory of the body is derived from the maximum distance between the edge boundary of the corner and the bottom of the robot when it travels in the convex environment. Additionally, we find the relation between kinematical structure of the robot and its ability of avoiding collision. The relation is realized by considering the workspace and the best posture of the robot in the convex structure. To provide necessary information for the algorithm, we use an IR sensor attached in the leg of the robot to perceive the convex environment. The validity of the gait planning algorithm is verified through simulations and the performance is demonstrated using a quadruped walking robot, called "MRWALLSPECT III"( Multifunctional Robot for WALL inSPECTion version 3).