Subject:
|
Re: Mission 1 giving lego robots power of intelligence and memory form Dr Patel
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 3 Aug 2000 17:04:12 GMT
|
Original-From:
|
Marco C. <marco@soporcel{ihatespam}.pt>
|
Viewed:
|
801 times
|
| |
|
|
hmmm... Have you ever thought of browsing the
http://www.lugnet.com/robotics section and read some of the discussions
*already* taking place about similar problems together with various already
available solutions and hipothesys ?
First (not second) problem: Moving. Only in a very controlled environment
(a special maze) will exist a 99% acuracy to redo a path exactly the way it
was done before. Even the task of reversing the steps made to get back to
the original starting point might still have some problems.
I remember doing a maze solving program when I was 17, that was perfect...
while in a computer screen "environment". One of the things real-life
(er... "toy-life" ;) robotics do for a programmer, is exactly introduce the
"unexpected" inacuracy in sensors, motors, dealing with 3D surfaces that
also induce errors in "2D" odometry, etc...
(Any fuzzy logic expert here besides Bert van Dam - www.bvandam.net ? ;)
So, I don't know if refining a Mapping algorithm based *only* in a
controlled environment through odometry and/or timerecording and/or
touchsensing should be the first task (mission?) to solve because it will
fail when you get to the other more in-real-life tasks/problems.
Anyway, just thought of sending this message to simply remeber you that
this is one of many topics that are (and were) *already beeing discussed*
in this list.
That's MHO.
____________________
Marco C. aka McViper
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
