Author Topic: Maze solving robot  (Read 683 times)

0 Members and 1 Guest are viewing this topic.

Offline IbaeniTopic starter

  • Beginner
  • *
  • Posts: 4
  • Helpful? 0
Maze solving robot
« on: January 25, 2014, 09:49:17 AM »
Hey guys. I'm junior and am working on a project for TSA. I have to build a robot that can be controlled (RC, bluetooth, etc.) and can run autonomously. The robot is to solve a maze with a solid black line centered along the maze. It seems as if there are only one or two T turns or 4 way turns. The robot will have no previous knowledge of the maze beforehand. It seems to me the way to do this is a right/left hand method. The event is with middle schoolers as well which is why I think it's this simple. I was wondering if anyone could give me any ideas as to a more advanced/intelligent method to run this course?

Currently my plans were either a simple push sensor for left hand method, an array of photoresistors for line following, or sharp ir to sense walls/gaps in walls.

Im not sure what other info to include. Ill answer any questions about the event that I can, but the instructions given to us by TSA are very vague.

Offline jwatte

  • Supreme Robot
  • *****
  • Posts: 1,343
  • Helpful? 81
Re: Maze solving robot
« Reply #1 on: January 25, 2014, 01:40:49 PM »
I would suggest the IR sensors if the black line is easy to detect.
Right-hand (or left-hand) following works as long as there are no cycles in the maze.

The only way to solve a maze better than that is to have global knowledge of the maze. You might use a tall pole with a camera on it, take a picture of the maze, do image processing to plan the path, and then follow it? :-)

Offline CJAlbertson

  • Jr. Member
  • **
  • Posts: 19
  • Helpful? 1
Re: Maze solving robot
« Reply #2 on: February 04, 2014, 02:29:47 AM »

Im not sure what other info to include. Ill answer any questions about the event that I can, but the instructions given to us by TSA are very vague.

If you can't know any details about the mass, like the width of the black line and color of the floor and how the room lights will be and soon  You have to use a left hand contact sensor.  Because you can't know of the maze walls will even work with the Sharp sensor.   Most sensors need to be tested with the actual surfaced they will work with.  But a microswitch and some spiny wire is bound to work on any wall.

I would work in this order
1) get the RC part to work.  This will allow you to test the motors and battery and so on
2) build an L/R hand sensor with simple switches, two per side so the robot keeps correct distance from wall
3) only after the above is reliable try to ad a smarter sensor but these can be fooled.  But you have #2 and even #1 as fall backs.
Chris A. in So. Cal.

Offline Webbot

  • Expert Roboticist
  • Supreme Robot
  • *****
  • Posts: 2,166
  • Helpful? 111
Re: Maze solving robot
« Reply #3 on: February 04, 2014, 05:58:08 PM »
If it's still within the rules then an old trick was to have a touch sensor that runs along the top of the walls. But the 'trick' is that it can also go slightly over the wall and thus sense other walls coming in at right angles - ie you can 'see through' the wall to the other side and hence build your map quicker.
Webbot Home: http://webbot.org.uk/
WebbotLib online docs: http://webbot.org.uk/WebbotLibDocs
If your in the neighbourhood: http://www.hovinghamspa.co.uk

 


Get Your Ad Here