통합검색
 
닫기
통합검색
 
  • home
  • Research
  • Current researches

Research

 


Development of Recognition Technology and Gripper Capable of Random Pieces Picking


 

- OVER VIEW -

  The all-in-one universal gripper integrates finger tips and suction grippers to grasp and manipulate objects of various shapes, sizes, and materials in industrial fields. To enable this, a vision system for determining an object is required. The focus is on increasing success rate and productivity through vision systems that estimate 6d pose and detect grasping points for a wide variety of random objects.

  This system can perform a universal gripper capable of multi-product picking, vision technology of random piece picking, grasping control algorithm using sensors, and high grasping success rate using multi-modal sensing.

 


All-in-one Universial Gripper


 

   It is possible to strategically grasp an object depending on the situation by utilizing a pneumatic gripper with a parallel/rotational joint and an electric gripper with an adaptive structure with multiple rotational joint. Accordingly, it is possible to grip and transfer an object more stably than when only one gripper is used. In addition, the object can be gripped regardless of the shape, size, or material of the object. 

 

 

6D Pose Estimation


 

  6D pose estimation detects features based on RGD-B images obtained from the camera. As this is classified through a pre-built learning model, 6D data of each object can be obtained. By building a model in advance for objects that can be easily seen in industrial sites, it is possible to perform the picking work immediately without the pre-registration process. Due to these advantages, the learning-based recognition system is suitable for industrial sites that are currently changing into multi-product.

 


Detection of Grasping Point


 

   It receives RGB and Depth images of several objects scattered within the region of interest from the camera. This detects feature points through feature backbone and detects grasping points for each object based on the neural network learning model. For each object, a point with a high grasping success rate is obtained as a result, and the gripper grasps that point.

 


 


Publications


  •  
  • • Seunghwan Um, Heeyeon Jeong, Chun Soo Kim, Issac Rhee, Hyouk Ryeol Choi"ReC-Gripper: A Reconfigurable Combined Suction and Fingered Gripper for Various Logistics Picking and Stowing Tasks", IEEE Robotics and Automation Letters, vol. 9, no. 1, pp. 87-94, Jan. 2024
  • • Yeong Gwang Son, Tat Hieu Bui, Juyong Hong, Yong Hyeon Kim, Seung Jae Moon, Chun Soo Kim, Issac Rhee, Hansol Kang, Hyouk Ryeol Choi"CoAS-Net: Context-Aware Suction Network With a Large-Scale Domain Randomized Synthetic Dataset", IEEE Robotics and Automation Letters, vol. 9, no. 1, pp. 827-834, Jan. 2024
  • • Issac Rhee, Gitae Kang, Seung Jae Moon, Yun Seok Choi, Hyouk Ryeol Choi, "Hybrid impedance and admittance control of robot manipulator with unknown environment", Intelligent Service Robotics(2022): 1-12.
  • • 손영광, Quang Huy Nguyen, Bui Tat Hieu, 문승재, 김춘수, 홍주용, 최혁렬,"랜덤피스 피킹을 위한 포인트 클라우드 기반 공압 그리퍼 파지점 검출 알고리즘",제 17회 한국로봇종합학술대회, 2022.05.11~ 05.14
  • • 김춘수, 이이삭, 최혁렬, "진공 그리퍼를 위한 2자유도 와이어 기반 다단 병진 및 회전 메커니즘 개발", 제 17회 한국로봇종합학술대회, 2022.05.11~ 05.14
  •  

Project


  •  
  • • 산업통상자원부 
    다품종 랜덤 피스 피킹이 가능한 인식기술 및 그리퍼 개발 (2022.5.16 - 2024.12.31) 

 


Researcher


Issac Rhee, Seung Jae Moon, Hee Yeon Jeong, Bui Tat Hieu, Chun Su Kim, Yeong Gwang Son, Seung Hwan Um, Ju Yong Hong, Yong Hyeon Kim