Kang, Seunghyun Lee, Jin hwi 서태원 Kim, Chang Hwan 2024-01-12T03:42:18Z 2024-01-12T03:42:18Z 2022-11-29 2022-05-12 https://pubs.kist.re.kr/handle/201004/77212 This paper presents a method to determine a grasp pose for object manipulation considering the configurations of robot. In such environments as shelves, a vision system and algorithm usually give an object’s information and multiple grasping plans. The grasping plan can also be given in a range of angle, but other objects may partially block the angle range. However, it is necessary to determine an angle range with a direction vector to control the manipulator, where a gripper can approach the target object. For this, we find a collision-free angle range by employing the modified Vector Field Histogram+ and determine an approaching angle that maximizes suggested manipulability to the target object. From the simulation results, the proposed method showed maximum 54% less planning attempts to determine a grasp pose and a approaching angle in clutter than others so that it could reduce the subtasks to search possible grasp poses mean by 36%. Korean 한국로봇학회 Manipulability 안에서의 방향벡터를 이용한 물체 파지 결정 방법 Determining a Grasp Pose Using a Direction Vector in Manipulability Ellipsoid Conference 2 제17회 한국로봇종합학술대회 제17회 한국로봇종합학술대회 KO 강원도 2022-05-11 제17회 한국로봇종합학술대회 논문집