Mapping Algorithm

Hello Group,
I have coded software that greatly enhances my robot to map and store coordinates of map in memory.
Here is the scheme:
START)go forward until bumper hit -sonar sweep 180 (SWEEP)derive equations of lines from data -see if lines are extensions of previos lines, if so, add length to previous line -pick a line that doesn't end in sensor sweep, go to spot on map where sensor sweep cut off last time. -goto sweep
The bot gets stuck sometimes, but I can simulate him on the graphics screen and it cuts down development.
The code is in BASIC. Anybody know a good spot to get information on smarter mapping programs, with EASILY understood concepts, code?
Rich aiiadict AT hotmail DOT com http://rich12345.tripod.com
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@hotmail.com (Rich J.) wrote in message
Rich, you might tale a look at the following - Where Am I? About 14MB in the download.
http://www-personal.engin.umich.edu/~johannb/position.htm
- dan michaels www.oricomtech.com =============================
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@hotmail.com (Rich J.) wrote in message
Hi,
I am using Lees algorithm for a "go to" command on the mapper. I was stuck on how to represent places already explored, but I've been playing text adventure games and I realized that I can map explored areas out just like text games!
Is there any code out there that shows how to use mapping algorithms on a real robot with sonar/bumber sensors (distance/touch).?
I have mine working very nicely right now. It was a bit difficult to deal with Lees Alg and variable room sizes.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
You might find some value in looking at occupancy grid algorithms. The following URL has some discussion of one such implementaiton and references that could help get you started.
http://rossum.sourceforge.net/papers/Localization /
I suspect that pretty soon some smart person is going to do for the robot navigation problem what Huffman did for information theory or what the Kalman filter did for estimation theory.
Gary
snipped-for-privacy@hotmail.com (Rich J.) wrote in message

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@hotmail.com (Rich J.) wrote in message

I have this working very well now! It uses a bunch of memory, and it is slow (interpreted basic, C in devel), but it works! No matter how complicated the maze, the robot can navigate it! It stores walls in memory as start/end points of lines. The C inplementation will use less memory. Basic uses full bytes for some things that only need 1 bit.
I had some trouble with doing searches and exploring and how to represent explored areas, and areas blocked by walls. It works now. Lots of fun! Still need to work some on inaccuracy of sensors (sonar, odometers). Sometimes the robot gets a little confused if it has traveled far, with lots of turns, and sensor readings get a little off. I have a subroutine to reposition the robot on the memory map if the readings get too far off.
Rich aiiadict AT hotmail DOT com http://rich12345.tripod.com
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.