Graphical User-Interface for the Shadow-Hand
Bachelor- or Master-Thesis at group TAMS
Motivation
The Shadow Dexterous Robot Hand is an advanced robotic hand
that closely matches the size of a male human hand.
With four joints per finger and five joints for the thumb,
and an actuated wrist, it can mimic a large subset of
human grasps and manipulation motions.
The software provided for the hand consists of a real-time loop
that generates motor commands and reads sensor values at 1kHz rate.
Integrated into the ROS framework, the rviz tool allows for
3D visualization for the current hand pose.
We also have a very basic user interface that allows to select
the hand control mode and to use graphical sliders to control
individual motors,
and software for hand teleoperation from a dataglove or from
RGB-D camera data.
However, we currently have no nice graphical user-interface
that would allow us to quickly switch between the different
control modes or to visualize the internal state of the hand
(e.g. joint positions and joint torques, motor temperatures,
tactile fingertip data, deviations of current hand state
from the last commanded values, errors, etc.).
Potential application scenarios
- robot hand control, calibration, and debugging
- graphical presentation of relevant internal robot state
- quick identification of hardware and software problems
- protecting the hand and users
- integration into existing user-interfaces (rqt/rviz)
The precise scope of the software implementation is to be discussed,
with a basic functionality achievable in a BSc-thesis,
while several of the more advanced topics are more in range of MSc-thesis.
Given the current Corona restrictions,
large parts of the software can be designed (and tested)
with the existing documentation and simulation models of the hand.
Actual evaluation of the software will probably require tests and
experiments on the real robot in our lab.
Thesis Goals:
- development of a graphical user-interface for an advanced robot system
- using the ROS open-source framework and environment
- efficient graphical presentation of internal robot state (2D and/or 3D)
- integration into existing safety mechanics (e.g. panic switches)
- going beyond joint-sliders: selection of grasp types and autonomous grasp execution,
generation of hand poses from inverse kinematics and inverse dynamics,
implementation of human-like finger synergies,
integration of tactile data into reflex-like motions, etc.
- filtering of relevant information (e.g. motor overload) from a huge stream
of input data
- etc.
- collect ideas and experiment data leading to a publication (MSc-thesis only)
Requirements
- as always, interest in the topic area, here graphical user-interfaces and robot manipulation
- programming skills in Python or C/C++ and Qt
- basic knowledge of the ROS framework
Contact
