Go Down

Topic: Is it possible to use compass sensor without using gps (Read 1 time) previous topic - next topic

chokey

I want to develop  a robot that can move in a maze without using any controller. My concern is,it is possible or not to use the compass sensor without using any gps?

Shpaget

#1
Mar 01, 2013, 10:43 am Last Edit: Mar 01, 2013, 11:06 am by Shpaget Reason: 1
Sure it is. Compass sensor is just a sensor like any other.

Have a look at the links listed here. They contain datasheet, tutorials, wiring diagrams and even some code.

edit:
Fixed the link.

Riva

Yes you can get compass sensors without having to resort to GPS. Maybe start here or on eBay, there are several different flavours and prices.
http://forum.arduino.cc/index.php?action=unread;boards=5,67,10,11,66,12,15,17,21,22,23,24,25,29;ALL


123Splat

O.K., you use a compass, but for what when solving your maze?   You don't need a compass to keep track of where you are going, or have gone.  You do that by storing distance traveled and direction change invoked by your obstacle avoidance routine (one from some senseo to count pulses generated by your motion (like an IR photo-interrupt), an the other from your code deciding what to do. 
So,,,, you got: I was traveling gorward, hit something, backup a little, turn left/right, zero counter, start moving forward...

Did I miss something, or am just plain screwy?  where does the compass come in?

PeterH


Did I miss something, or am just plain screwy?  where does the compass come in?


Knowing which way you are pointing seems like a very useful thing to know when you're trying to navigate a maze.
I only provide help via the forum - please do not contact me for private consultancy.

Shpaget

I think Splat is trying to say that it can be done by dead reckoning alone.

wildbill

It can indeed, but throw in some wheel slip and it would be really nice to have a compass too.

holmes4

Quote
I think Splat is trying to say that it can be done by dead reckoning alone.


Quote
Knowing which way you are pointing seems like a very useful thing to know when you're trying to navigate a maze



No!

The way to navigate ANY maze is to keep your left hand on the wall. And that's all!. Try it it always works.

You could record your movements and then trim the path to find the fastest way.

Mark

retrolefty

#9
Mar 02, 2013, 01:46 am Last Edit: Mar 02, 2013, 01:49 am by retrolefty Reason: 1
My wife says I shouldn't try and figure out how to get somewhere on my own, I should rather just ask someone for directions. Women, can't live with them, can't live without them.  ;)

I tell her that men are born with an internal compass, it's a hunter/gather thing from early evolution times, but she doesn't buy it.
Lefty

PeterH


Try it it always works.


Unless you're trying to solve a maze where that algorithm doesn't work, or where you're trying to find an optimal solution rather than just a solution.

You don't always need a compass to solve a maze, but there are some situations where it's a useful thing to have.
I only provide help via the forum - please do not contact me for private consultancy.

retrolefty



Try it it always works.


Unless you're trying to solve a maze where that algorithm doesn't work, or where you're trying to find an optimal solution rather than just a solution.

You don't always need a compass to solve a maze, but there are some situations where it's a useful thing to have.


Well the always turn left or always turn right method does prevent one from 'doubling back on oneself' and therefore getting into hopeless endless loops. But certainly it is a most inefficient algorithm to use and will never win any competitive events unless all the others happen to have mechanical breakdowns.  :)

holmes4

Quote
Unless you're trying to solve a maze where that algorithm doesn't work,


Do you have an example of such Peter, I'd like to see it.

Quote
or where you're trying to find an optimal solution


Which is why I said find the path then trim it ( removing all dead ends).

Mark

holmes4

Quote
But certainly it is a most inefficient algorithm to use and will never win any competitive events unless all the others happen to have mechanical breakdowns.


And a better algorithm is? (For a maze you have never seen before).

Mark

retrolefty


Quote
But certainly it is a most inefficient algorithm to use and will never win any competitive events unless all the others happen to have mechanical breakdowns.


And a better algorithm is? (For a maze you have never seen before).

Mark


I've never tried to design one, but if I were I think I would start by checking the web: Goggle states 1+ million hits on "maze algorithm", this seems to be a well travelled path.

http://en.wikipedia.org/wiki/Maze_solving_algorithm

http://www.astrolog.org/labyrnth/algrithm.htm

http://www.cs.bu.edu/teaching/alg/maze/

Lefty

Go Up