DPRG
DPRG List  



DPRG: UV navigation?

Subject: DPRG: UV navigation?
From: Cyborg droid at bellatlantic.net
Date: Thu Jul 31 16:47:21 CDT 1997

Well...the idea seems quite cool (especially that I have vacuuming the 
room.)

Hmm, you know, I have been watching the Discovery channel (a great source 
of the newest gadgets are Beyond 2000 and Next Step.) and what they have 
shown one time was an robo-mower. This thing was powered by the solar 
baatteries, could climb %10 hills and I believe that it had a global 
positioning system on it(with resolution to like few inches or 
something).The point is that that robot was expensive as hell. The area 
of teh lawn was programmed to the robot and then you could let it go 
alone and cut the lawn. Well...I woudn't recommend that since your 
neighbor would most likely steal it or brake it(well, maybe not).
Probably using the Basic Stamp 2 is not such a bad idea. There is a 
problem. Like you said , there would be a need of 'plugging' the cable in 
the floor. 

What I thought is maybe more complicated but at the same time more 
challenging and also a better idea. But before I start, there is few 
details I have to talk about. 

You all know that human uses a 3 dumensional system to determine where 
the things are (for the need of storing them in the memory in case you 
want to return to the place or walk among 'obstacles')

Most of the robots that use IR know where the obstacles are simply by  
sending the signal from one of FEW emitters and then by receiving the 
signal, it knows where the obstacle was(it simply remembers,on which side 
the signal was received)

I haven't looked at the systems which use camera and how they work ( I 
will tough,but later), but sonars seem to be very useful.

Now, another thing is that these ssytems are not exact enough.
I don't know if any of you guys watch the Discovery channel's "Movie 
Magic" series. In one episode they have shown how the face of the 
Terminator 2 was scanned into the computer. The way they have done it was 
 as follows.
They have placed the actor in front of the laser beam which have moved 
around his head. This way a 3D image was scanned into the computer.

Now I know that it would be rather complicated to do BUT there is how I 
see it:

1) Make a laser beam gun on the rotatable base.
2) Make the gun shoot the beam vertically and then horizontally (all  
matter of the mechanical gizmoes)
3) Store in the memory the distance form the obcejts vertically and 
horizontally ( the resolution of the map would be to the nearest 
centimeter [1 inch = 2.54 centimeter])
4) Link these these parameters to make 3D map. Then the computer woould 
actually 'remember' the 'plan' of the room almost as the human does.
5) Program the robot to move among the obstacles an dvacuum the floor.


The positive points :
You don't have to rewire the room every time you change the placement of 
the furnitures. it scans teh room for the obstacles every time before 
vacuming.

The negativ points :
price,
and how to make this sucker work when there is more obstacles in the room
say, the robot cannot see what is behind the chair that it is scanning.

Possible solutions :

Make robot scan the room from 4 different corners of the room ?

Please send any suggestions.

Truly yours ,
Cyborg.

------------------------------

More information about the DPRG mailing list