Modular robotic hand with vision object recognition capability

The main idea of this project consists of developing a modular robotic hand. A camera is embedded in the hand palm. This will allow for a stereo vision distance estimation of the object to grasp using only one camera. The idea consists in taking two pictures from different point of view by moving the arm where the hand will be placed. The two photos are then used to create a stereo image in order to estimate the distance from the object to grasp.

Possible approach

In order to realise the grasping functions, Force Sensitive Resistors or servo current draw measurements can be used. The second approach allows for more accurate measurements of the torque which each servo is executing during the grasping phase.

Current implementation

Students: Sund Kim Andrè, Thorsen Kjetil, Stangeland Svein Rune, Godfrey Brita Erica. Supervisors: Filippo Sanfilippo, Houxiang Zhang.

It should be noted that the stereo vision function has not been implemented yet.

​

This student project constitutes the preliminary study behind one of our newest and open source projects which is ModGrasp.

Future works

• Integration with the KUKA KR6 R900 Agilus robotic arm.
• Stereo vision distance estimation.
• Brain control of the robot by using an mind wave headset