Society of Robots - Robot Forum

General Misc => Misc => Topic started by: dannytemmerman on February 24, 2009, 09:04:33 AM

Title: Mapping robot
Post by: dannytemmerman on February 24, 2009, 09:04:33 AM
Would it be possible to make a mapping robot with a ATmega168 from the $50 robot and a couple of ultrasonic sensors? And what kind of program should I upload? Because I haven't got a clue what program I should need.




-dannytemmerman
Title: Re: Mapping robot
Post by: HyperNerd on February 24, 2009, 09:55:48 AM
Elaborate further ;)!

Do you mean for the robot to be placed at the start of a maze, and drive around recording where the walls are? If so, then yes, this should be quite easily accomplished. Maybe the robot could use a wavefront algorith (http://www.societyofrobots.com/programming_wavefront.shtml) to work out where it needs to be.

Also, how about staring with bog-standard maze solving code, and adapt it so it records where it finds walls. There are so many ways you could do this, that it would really help if you expanded your idea :)

 -HyperNerd
Title: Re: Mapping robot
Post by: dannytemmerman on February 24, 2009, 11:54:37 AM
Elaborate further ;)!

Do you mean for the robot to be placed at the start of a maze, and drive around recording where the walls are? If so, then yes, this should be quite easily accomplished. Maybe the robot could use a wavefront algorith (http://www.societyofrobots.com/programming_wavefront.shtml) to work out where it needs to be.

Also, how about staring with bog-standard maze solving code, and adapt it so it records where it finds walls. There are so many ways you could do this, that it would really help if you expanded your idea :)

 -HyperNerd

No, I mean in a room. The robot goes to one determent space and than goes back. I is this hard to do in code?
Title: Re: Mapping robot
Post by: superchiku on February 24, 2009, 12:14:00 PM
no mapping and path finding algo's have to be used...use vision
Title: Re: Mapping robot
Post by: dannytemmerman on February 24, 2009, 12:15:53 PM
no mapping and path finding algo's have to be used...use vision

Vision is to expensive. Isn't it possible with ultrasonic?
Title: Re: Mapping robot
Post by: HyperNerd on February 24, 2009, 02:02:10 PM
How big is the room?

Ultrasonic becomes less accurate at distance, so you may need to rethink :(

Have a look at this: NorthStar Indoor 'GPS' (http://www.evolution.com/products/northstar/)
That might prove to be what you need

 -HyperNerd
Title: Re: Mapping robot
Post by: HDL_CinC_Dragon on February 24, 2009, 02:37:03 PM
1) I dont think thats what hes asking for. I think he wants to be able to store a map of the environment. From what I read on that site, it just uses localization using IR to know where it is relative to the IR transmitter.

2) its $1800 lol
Title: Re: Mapping robot
Post by: superchiku on February 24, 2009, 02:48:39 PM
vision isnt expensive all u need is a camera and a pc...directly control the motors from the pc...ill teach u how to if u want...
Title: Re: Mapping robot
Post by: HyperNerd on February 25, 2009, 01:50:28 AM
I didn't realise it was that expensive :o

If you use stepper motors, have the robot run round the edge of the room, and then spiral inwards, recording where it finds obstacles.

I say stepper motors as they are extremely accurate

 -HyperNerd
Title: Re: Mapping robot
Post by: superchiku on February 25, 2009, 02:03:27 AM
asi told u ..if u use normal scanning ir's and all ...ur mapping algo has to be very cool...but what if the obstacle changes its position...then u need vision my dear...the logic which u say is ok ..if u know the dimensions of the room...but u need to use  a lot of hit and trial...for it...
Title: Re: Mapping robot
Post by: Admin on March 02, 2009, 12:54:45 AM
Mapping with ATmega168 . . . just replace the Sharp IR with Sonar:

http://www.societyofrobots.com/robot_iRobot_Create_mod.shtml (http://www.societyofrobots.com/robot_iRobot_Create_mod.shtml)