Grasping Known Objects With Aldebaran Nao Mohd Dawood

Grasping Known Objects With Aldebaran Nao Mohd. Dawood Ashu Gupta Mentor-Dr. Amitabha Mukherjee

Importance of what we are doing � Partcipating � Using in Robocup. humanoids for trivial day-to-activities.

Ball Detection Take Input Image Convert To HSV Image Create Binary Image

Ball Detection Binary Image Smooth Binary Image Detect Circle , center and radius

Training for Calculation of Position of ball �

� Radius and Distance Values. r=23 Depth=8. 3 cm Error=± 5 mm r=28 Depth = 4. 2 cm Error = ± 3 mm r=33 Depth = 2. 1 Error = ± 3 mm

Kinematics of Nao We will be using only the left hand for grasping. *Image taken from Nao's Documentation

Reaching to the Ball � Solving the inverse kinematics equation for the arm we get θ 1 and θ 2 and then we move arm such that the end effector reaches P 3 and the arm angles are θ 1 and θ 2.

Video

Future Improvement � Grasping � Visual of object without fixing it’s height. Servoing. � Implementation more accuracy. of Virtual Thumb Rule for

References � Master’s Thesis of Tomas Gonzalez Sanchez, Depatment of Computer Science and Mathematics, Universitat Rovira I Virgili � Documentation of Aldebaran Nao from Aldebaran Robotics. � Website of Aldebaran Robotics http: //www. aldebaran-robotics. com/