Subject:
|
Re: Gus, congratulations
|
Newsgroups:
|
lugnet.org.us.smart
|
Date:
|
Thu, 3 May 2001 19:42:33 GMT
|
Viewed:
|
1092 times
|
| |
|
|
In lugnet.org.us.smart, Dave Fisher writes:
> Can't believe no one has mentioned it yet. Gus took his Lego robot to the
> SRS Robothon last weekend and took FIRST PLACE in the Line Maze. That
> deserves a good round of applause. You have set a standard for us all to try
> and achieve. Well done.
Thank you Dave, (applause), Thank you, thank you everyone (more applause)...
Actually, SMART members David Schilling and Barton Place were also at the
Robothon and have already congratulated me.
The speech:
The Line Maze was a very challenging event. As far as construction goes,
the trick was to figure out a way to arrange the light sensors so that they
can accurately track a 1/4" line as well as reliably detect intersections.
But compared to programming the robot, building it was trivial. This was by
far the most complicated NQC program that I have written.
I'm working on a LUGNET page about Alegomazer, the name of my robot, which
will go into more detail about how it works. I will post a link when I have
that done.
But to wet your appetite, I will give you a bit of a summary.
In order to backtrack and learn from previous runs, the program implements a
stack. The stack uses only 16 integers of the 32 global variables
available, to store as many as 45 intersections, considerably more then the
6 or so that were actually required at the Robothon maze.
The robot runs a specified amount of time (officially 3 minutes at the
Robothon) and then stops. On subsequent runs, the robot knows which paths
it has already tried and which paths it still needs to explore. If the
robot has already solved the maze on a previous run, it runs straight to the
finish without any wrong turns/dead ends.
Because the robot keeps multiple light threshold variables and other maze
navigation data, the remaining 16 global variables were not sufficient. To
give myself more variable space, I used the upper, lower, and hysteresis
thresholds of unused events. This scheme works great, though there are some
important limitations you should know if you plan to try it. I will talk
more about it in my Alegomazer page.
The official rules and a sample maze can be found here:
http://www.seattlerobotics.org/Robothon2001/linemaze.html
Some Alegomazer photos, as well as photos of the 2nd place LineMax and 3rd
place Dilbert II, can be found at:
www.brickshelf.com/gallery/GJansson/Alegomazer
Gus
|
|
Message is in Reply To:
 | | Gus, congratulations
|
| Can't believe no one has mentioned it yet. Gus took his Lego robot to the SRS Robothon last weekend and took FIRST PLACE in the Line Maze. That deserves a good round of applause. You have set a standard for us all to try and achieve. Well done. (24 years ago, 2-May-01, to lugnet.org.us.smart)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
