Controlling a Robot Hand using Visual Feedback
External reference persons Marco INDACO (Post Doc)
Research Groups TESTGROUP - TESTGROUP
Thesis type EXPERIMENTAL
Description Within the framework of the PARLOMA project (www.parloma.com) we are developing a 3D Printed low cost anthropomorphic robot hand for human-robot interaction.
The problem of status estimation for feedback control is well known in all automation areas. It is one of the main issue when developing novel and integrated systems where feedback estimation from standard sensors is impossible due to system integration or costs.
The thesis proposal aims at developing a system able to estimate pose and configuration of a anthropomorphic robot hand using visual sensor, e.g., RGBD-cameras (as kinect) or standard cameras. The basic idea is based on the use of an external visual sensor to estimate finger position and handshapes in order to improve the control of the hand.
In particular, the system should be able to
1. Recognize the robotic hand within image view and segment the hand from the background
2. Estimate the joints position of the robotic hand
3. Use the joints estimation to design a position control system for robot hand.
Deadline 02/12/2015 PROPONI LA TUA CANDIDATURA