Boe-bot maze memorization

kingbp

I have a project with the Boe-bot where i want to make it drive through an entire maze with dead ends and continue until it reaches the end of the maze on the first run. Then It will make a 2nd run this time memorizing the entire maze and taking the shortest route to the end of the maze avoiding dead ends. Can any of you guys help me with the p-basic coding. Im new to basic coding and am having trouble with this. Thanks!

erco

Line following or wall following? What types of sensors? Post your code so far and let us know what you have achieved already. Lots of maze-solving algorithms out there, and yes they can be very tricky to implement.

Check the accelerations on the final run of this 3pi robot: http://www.youtube.com/watch?v=mJV-KDqHgDQ

ajward

erco wrote: »

Check the accelerations on the final run of this 3pi robot: http://www.youtube.com/watch?v=mJV-KDqHgDQ

Now that's pretty damn cool!

Amanda

Humanoido

You could data log the turns to get through the maze the first time, then recall the sequence on the trip back.

Martin_H

If I was a beginner I'd go for a line maze over a maze with walls. Following a line and determining junctions and so forth is much simpler than interpreting sensor data to locate walls. Once I made that decision I'd break it down into steps. First learn how to follow a line, then learn how to determine that I'm at a junction.

Here's a link in the old forums on using the QTI sensor to solve a maze:

http://forums.parallaxinc.com/forums/default.aspx?f=6&m=341331

Here's a link to a PDF on line maze solving from the other guys:

http://www.pololu.com/file/0J195/line-maze-algorithm.pdf