Research about Faraman Robot and Dexterous Manipulation


 

 

- OVER VIEW -

  As we know, manipulation and exploration go hand in hand. Pure manipulation occurs when the object is completely known and pure exploration happens when the object is fixture and is not known. Dexterous manipulation is a combination of the two, we need manipulation for exploration and vice versa. It is challenging to integrate tracking and shape reconstruction into dynamic manipulation. How to reconstruct the shape of the unknown object during the manipulation is a hot topic today.

 


Exploration of Global Surface Geometry Local Information


 

The global surface of an object plays important role in achieving stable grasp and manipulation using a robot hand. In this research, we propose a method to explore global shape of an unknown object from local information. 

 


Faraman Robot


 

Faraman AS1 is an articulated robot with 6 DOF of Samsung company. We use it as the manipulator to get the contact with the object. Before control we calculate some mathematics problems: forward kinematic, inverse kinematic and Jacobian matrix. We use PI controller to control the position of the robot. Some trajectory control algorithms have been done in Cartesian space such as: straight line, rectangle and circle motion. 

 


Detect contact position using 6-axis Force Torque sensor


 

The problem is we already know the surface equation of the sensor tip, the measured force and moment from 6-axis force torque sensor, how we can calculate the contact position or contact centroid between the sensor tip and the object. We use two methods depending on the contact between sensor tip and object. If the contact is one point, we use the point contact with friction method. If the contact is a small area, we use the soft finger contact method. The second method is usually used in the research. 

 


Estimating Principal Curvatures


 

The local shape of the object is characterized with the curvature. However, the number of curvature at a point is infinite in 3D shape. The normal curvature of an object along any direction can be computed when we have two principle curvatures. In our research we build the algorithm to calculate the two principal curvatures using 6-axis F/T sensor. Three normal curvatures along 3 known directions are calculated from the measurement and from that we can estimate the principal curvatures using above equations. 

 


Local Shape Reconstruction with Minimum Number of Contact Points


 

In this research, we explore the local shape of unknown object using 6-axis Force Torque sensor and a 6-DOF Manipulator, FARAMAN robot. We suggest the method to reconstruct the local shape of unknown object by combining contact position and normal vector information. Our method just need 4 contact points, smaller than previous methods, to estimate local shape of an object. Experiments have been done with several objects to verify the proposed algorithm. 

 


Publications


 

  • • Tri Cong Phung, Min Jeong Kim, Hangil Park, Dongmin Choi, Seung Hoon Shin, Hyungpil Moon, Ja Choon Koo and Hyouk Ryeol Choi, "Exploration of Global Surface Geometry from Local Information", 2012 9th International Conference on Ubiquitous Robots and Ambient Intelligence(URAI), Daejeon,Korea/November 26-29,2012
  • • Tri Cong Phung, Min Jeong Kim, Hyungpil Moon, Ja Choon Koo, and Hyouk Ryeol Choi,"Exploration of Local Surface Geometry with Minimum Number of Contact Points and Surface Normal Information",International Journal of Control, Automation, and Systems, Vol.10,No.2, pp.383-395, 2012/4
  • • Tri Cong Phung, Hansang Chae, Min Jeong Kim, Dongmin Choi, Seung Hoon Shin, Hyungpil Moon, Ja Choon Koo, and Hyouk Ryeol Choi, "Estimation of Unknown Curvature using a Coarse-Resolution Sensor and Contact Kinematics", 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, September 25-30, 2011. San Francisco, CA, USA
  • • Tri Cong Phung, Min Jeong Kim, Hyungpil Moon, Ja Choon Koo, and Hyouk Ryeol Choi,"Exploration of Local Surface Geometry with Minimum Number of Contact Points and Surface Normal Information",International Journal of Control, Automation, and Systems, Vol.10,No.2, pp.383-395, 2012/4.
  • • Tri Cong Phung, Hansang Chae, Min Jeong Kim, Dongmin Choi, Seung Hoon Shin, Hyungpil Moon, Ja Choon Koo, and Hyouk Ryeol Choi, "Estimation of Unknown Curvature using a Coarse-Resolution Sensor and Contact Kinematics", 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, September 25-30, 2011. San Francisco, CA, USA
  • • Tri Cong Phung, Sungmoon Jin, Han Sang Chae, Hyungpil Moon, Ja Choon Koo, and Hyouk Ryeol Choi, "Reconstruction of 3D Local Surface Geometry by Using Minimum Contact Information", The 5th Int. Conf. on the Advanced Mechatronics (ICAM2010), pp. 450-455, 2010/10/04-2010/10/06, Japan.
  • • Tri Cong Phung, Yong Seok Ihn, Ja Choon Koo, and Hyouk Ryeol Choi, "An Enhanced Edge Tracking Method Using a Low Resolution Tactile Sensor", International Journal of Control, Automation, and Systems, Vol.8, No.2, pp.462-467, 2010/04, Korea
  • • Tri Cong Phung, Yong Seok Ihn, Ja Choon Koo and Hyouk Ryeol Choi, "Edge Identification of a Small Object through a low-resolution tactile sensor array", International Journal of Precision Engineering and Manufacturing, Vol.11, No.2, pp.247-254, 2010/04, Korea
  • • Tri Cong Phung, Byung June Choi, Seung Hwa Ha, Yong Seok Ihn, Ja Choon Koo and Hyouk Ryeol Choi, "Estimation of Local Shape and Pose of Convex Object via Active Touch", 한국지능로봇 하계 종합 학술대회, pp.267-268, 2008/06/25-2008/06/27.

 


Project


 

  • • 한국산업기술진흥원 
    스마트 로봇핸드를 이용한 로봇매니퓰레이션기술 전문인력양성사업 (2008.9.1 - 2011.8.31) 
  • • 지식경제부 
    스마트 로봇핸드를 이용한 로봇매니퓰레이션기술 전문인력양성사업 (2008. 9. 1 - 2013. 2.28)

 


Researcher


 

Feng Yi Liu, Mina Choi, Gi Tae Kang