Robotics this weekend?
Guy
have summer weather it's getting even harder to progress on the Robo-
Magellan entry.
I have what is hopefully my last parts order coming in by post which
doesn't appear to be arriving this week so hopefully next. It
includes a number of things that I want to add to the base platform
including current and IR sensors which will give me better and safer
motion control. Of lesser importance I'd like to hook up an audio
amplifier (I have a pair of LM380N-8 chips), hook up some of the
lights on the body shell, mount the forward sonar sensors better, and
try to reduce the sticking of the bumper. That's several weekends
right there.
In the meantime I'm working on the logic for the navigation unit which
is to be mounted on the top of the vehicle and will handle mapping and
vision. The plan is that the navigator generate a short sequence of 2
or 3 path segments that the pilot in the base will try to follow. In
addition the navigator continually updates the maximum speed,
acceleration, and distance values. From these the pilot uses it's own
sensors (range-to-obstacles, roughness, tilt, slippage) to determine
safe speed and acceleration values within those maximums and that will
bring the vehicle to a stop at the maximum distance.
We have 3 different platforms in the team and we seem to be working
totally independently. Is there some way that we can share the work
and improve our chances of success?
I'm off to Kamloops for the long weekend so this weekend is the last
before the next robotics club meeting. Anyone interested in meeting
and hammering out a common architecture at least?
Take care.
Guy
Jonathan Hanna
Yeah, the summer is interfering a bit, I have a guest visiting from
Sunday ....
Saturday afternoon I am around.
I have not added anything but a bit of orange seeking vision lately, mostly
just playing and experimenting with the CPU requirements of some
algorithms. I have not even started looking at anything like a pilot.
I am sure we could share something, but we certainly need more time
and coordination.
Tim Snider
No real progress here, and I'm booked today.
Cheers,
TJ
Guy Lancaster
through a rough architecture and sketched some major interfaces. From
the top we start with a mission map that's loaded into the Navigator.
My suggestion is that the Navigator be a separate component that we
can transfer between platforms and/or duplicate when we have a winning
one. The Navigator uses the map along with vision and global sensor
inputs to plan paths and track progress.
The Navigator frequently passes the next 2 or 3 path segments to the
Pilot (aka Driver) which does it's best to safely follow the path. It
replies back with the path that it's actually following which is
modified according to it's own local sensor feedback and safety logic.
I'll get the details of the map entries and the path segments written
up when I get a chance, possibly before the weekend. If I haven't
made some progress before the next club meeting I'm thinking that I
might skip the meeting and get some work done. The weekends are
skipping by too fast! :-)
Take care.
Guy