Henk, Angelo, & rest of the Space Farm Group,
I have looked into some steps to move forward with robotic arm
manipulation and artificial intelligence control. From last
conversations I had with Angelo, the goal was to produce an interface
that would allow a person to make high level planning decisions for
the robot arm and essentially create a task list for the day.
One such method successfully implemented in the sensor and robotics
lab I am studying at is a 3D object recognition with situation
assessment utilized through fuzzy cognitive maps. Effectively, it
allows a robot to look at and identify objects it is trained to see.
Plan out its arm trajectory to pick up the object and move it, and at
the same time correct for any deficiencies in that movement.
(Predictor-corrector)
A video showing this in action can be seen in youtube link. Please
forgive the god awful robot arm used by Martin, as it is is the only
one in our lab at the moment, and suffers from power drops from rather
primitive servos.
Despite the difficult problems with this equipment, Martin was still
able to get it working, by locating and picking up the cup and moving
to a new location:
http://www.youtube.com/watch?v=xUIEv0LdAiw&feature=youtu.be
I strongly believe that the ARM 2.0 will work extremely well under
this system. The servos on ARM 2.0 are sophisticated enough and the
arm itself is light years ahead of what you see in the video.
Moving forward, this will be a complex and time consuming task that I
think may be best served under a PhD/Master student grant with some
additional funding to hire Ileana during the summer months while she
works on her undergrad.
A proposal for this could be submitted through National Science and
Engineering Research Canada (NSERC) through the lab here under Dr.
Leung. There is experience in this lab along the correct lines. Also,
with a strong link to HI-SEAS, we expect it will have a strong chance.
In writing this grant we would ask for four years funding for:
- 1x PhD student
- 2x MSc students (staggered one after the other)
- 2x Summer Interns for four years (Ileana and other)
- Equipment Costs
- Travel Costs
Ideally we would have Martin Walker as the PhD student, since he is
already familiar with the starting point of the system. However, no
reason someone else couldn't take on the position.
This will be a fairly ambitious proposal, but we feel it has a strong chance.
To proceed, I guess I need to know if this fits in with the direction
that the space farm is going to move in. If there are going to be any
drastic changes to the system, then this particular method may no
longer be appropriate.
Henk, I assume you will continue to host the space farm at your ranch?
Or do you see it moving up to the habitat location?
Please let me know if this is a good way to move forward with your
current plans or not. We can then start hammering out a grant
proposal.
Best,
Simon