1. Home
  2. General Robotics Forum
  3. 3 point triangulation

3 point triangulation

I'm working on a project where the bot will be confined to a set walled area. I'm using shaft encoders to track where the bot is roaming about the play area (roughly 15 ft by 10 ft). In regards to errors that will ultimately compound as well as things like someone picking it up and moving it, where he is at power up, etc I also want to triangulate the location to sort of reset where the bot is. I was intrigued by a method in the Robot Builder's Bonanza using sonar coupled with IR light to triangulate a position (using 3 beacons). There wasn't a whole lot of info in the book about it and I was wondering if someone has attempted this or if there is some more info out there about it (parts, circuits, notes, etc)?

What might be some other methods to figure a dead spot position? I haven't explored GPS much, is this a viable option (using a Basic Stamp 2) and considering it's a relatively small bot on a 9 volt?

thanks

Reply to
teranews
Loading thread data ...

location to

The easiest solution for a robot with shaft encoders is maybe to reset itself by hitting the wall as is done by a commercial robot. It aligns with one wall and then with the other.

You don't say what the actual situation is as regards freedom to structure the environment? For example if you used a tiled lino floor and a simple sensor used in line following robots you could adjust the x or y coordinate everytime the robot went over a line?

JC

Reply to
JGCASEY

Thanks!

The environment will be basically a wooden tabletop.

The bigger picture: The bot will be controlled by a PC via wireless communication. The PC knows the layout of the area and will tell the bot the proper driving commands to reach his destination, but the PC needs to calibrate the bots position (and heading using a compass). The bot should be able to get to his position using his shaft encoders, but errors will occur, so when the bot "thinks" he has arrived, I would like for him to get a fix on his position, reply back to the PC so the PC can then give him driving instructions if he is off. This means, the bot should remain in place to get a fix on his position (although he can spin if necessary). GPS I believe would be the best solution, but it is costly and I don't know how accurate it reports given such a small area. The idea in the Bonanza book would be perfect since it triangulates his position while he is stopped. There will be other obstacles and bots living in the play area too (of which the bot will use detection to avoid). If the bot was to strike an obstacle, the PC would then need to find him again.

Reply to
teranews

necessary). GPS

stopped.

formatting link
In the Sep1999 newsletter there is a triangulation used for navigation.

Perhaps a using a spinning mirror 45 degrees to the horizontal and a laser with phototransistors around the table to pick up the beam pulse and send it to the PC for navigation?

Reply to
JGCASEY

Thanks again! Looks like a good article. I'll read through it.

Reply to
teranews

You can rule out GPS. Unless you do something even more expensive than conventional GPS, you won't even be able to tell if the robot is on the table.

Mitch

Reply to
Mitch Berkson

Ah, ok. I had read up on GPS in the mean time and did see it isn't really geared to something this small.

thanks

Reply to
teranews

There is a chapter on beacon nav/triangulation systems here - isa BIGGGG download ...

formatting link
dan michaels
formatting link
=====================

Reply to
dan

Also, check out the "NorthStar" system by Evolution Robotics. It uses IR light spot triangulation to obtain ground-truth data for indoor navigation and localization. Not sure if it's within your budget but it's definitely worth looking into. :)

formatting link
luck with your project.

- PJ

Reply to
PJ

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.