Lab exercises

Afternoon sessions will be for hands-on lab exercises, i.e. mini-projects. 6 exercises are prepared, with only 5 two-hour slots to do them. Attendees will be divided randomly into 6 groups and given a schedule of exercises – see below. Since five time-slots are scheduled, and 6 exercises organized, one is optional and scheduled for free-time on Thursday. Of course, all lab exercises are more-or-less optional.

Note that the groups are quite large, because of limited resources.

Distribution into groups is listed here.*

*Note, you can of course swap groups – exchange with someone so that the groups stay the same size. We suggest that you do exercises with people who are not from your institutions. If you are in the same group with people from your institution (because of random distribution), do change the groups with someone.

Some exercises will require the use of your own laptops.

Working on the robots will be possible in the free time between the morning- and afternoon- sessions, but with only minimal attendance of exercise tutors.

Six lab exercises are planned:

1. End2end vision to motion for flattening on Franka FR3
2. Vision-action-language (VLA) model application on Franka FR3
3. Data collection using a bimanual Franka FR3 system with Gello haptic interface (instructions at the station)
4. Mobile manipulation with PAL’s TIAGo Pro
5. Textile benchmarking with ABB’s GoFa
6. Deep reinforcement learning in simulation (on JSI’s cloud infrastructure – you should have received an email about requirements for this exercise)

The timetable for lab exercises is as follows:

Please check back for distribution into groups.