Aug 27, 2012 , by
Taks 7: System testing
Within our last experiment steps we evaluated and optimized the entire robot system. In the first step we examined the object recognition and position estimation of Camera A. The camera configuration and image processing were improved until the derivation did not left a range of ±50mm. In the next step we examined the positioning of the gripper. The system was improved until the deviation may amount to only ±5mm, so that the pick procedure can be done successfully.
After the evaluation of the cameras and image processing, we optimized the position control of the robot to improve the repeatability. After optimization we arranged the scenario as it was introduced in the proposal with a set of three different tagged cubes. In several repetitions we diversified the position of the cube in the range of Camera A, in order to determine the reliability of the system. The experiment was considered as final, as soon as the system was able to grip tagged objects in its vision range and deposit these objects in a defined target position.
Furthermore we arranged a second scenario in that small cylinders with a through-hole were threaded onto a needle. Here, the dimension of a single cylinder was 10x5mm (height by diameter) and the mentioned through-hole had a diameter of 3mm. The robot was able to thread all cylinders onto the needle, which verified a repeatability of ±1.5mm.
Taks 8: Dissemination
In the last two experiment month, the most important dissemination activities were:
- The Multimeda Report was created and is now available here: http://www.youtube.com/watch?v=o94NYrFNRmo
- We preparing the last presentation about our ALEXA experiment that will be given at the IROS 2012 during the ECHORD workshop. The workshop schedule is available here: http://echord.info/wikis/website/iros-2012-workshop
Photographs of the ALEXA robotic arm