Subject:
|
RE: IR scanner
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Tue, 30 Nov 1999 18:45:29 GMT
|
Original-From:
|
Jim Thomas <JIM.THOMAS@antispamTRW.COM>
|
Viewed:
|
630 times
|
| |
|
|
The biggest problem I see with using rotating platforms, etc. is that they
are slow and use up a bunch of I/O ports to implement well. There are a
number of "one pixel" sensors that you can use with these implementations --
several have been mentioned here. I am kindof captured by the triangulation
idea. Sound travels at around 1Kft/sec. So if you had three sensors
mounted on the robot no more than a foot apart I think you could get a very
good directional measurement from an ultrasonic audio beacon. You end up
with up to 1ms delay so you could make a hardware timer running on a 1Mhz
clock or less and still get plenty of resolution. If you had three
distinguishable beacons (different freqs) you could triangulate your
absolute position in the room. Not a trivial piece of hardware but you
could multiplex the delay data onto one analog using the mode switching
trick. The RCX could take the 9 delay times and do the appropriate calcs.
JT
> -----Original Message-----
> From: Kornel Lambert [mailto:kornel@black-hole.co.uk]
> Sent: Tuesday, November 30, 1999 8:39 AM
> To: lego-robotics@crynwr.com
> Subject: Re: IR scanner
>
>
> Jim Thomas <lego-robotics@crynwr.com> wrote in message
> news:BFBCB00A4D7BD211A9BB0000D11BAFAE26A829@mbsp05.sp.TRW.COM...
> > There is the classic Polaroid ultrasonic range finder. I
> think someone
> here
> > was working on one of these or similar for lego? They work on the • reflected
> > signal approach. Now I thought you wanted to do off board • triangulation.
> > You can accomplish that using a omni-directional source on • the robot and
> > three sensors around the room hooked up to your PC. You • might need custom
> > acquisition hardware or else the PC might have a ton of data to wade
> > through.
>
> How about several (>= 3) light sources in an otherwise dimly
> lit room, and
> the robot tracking them with a light sensor mounted on a tower whose
> rotation is measured by a rotational sensor?
>
> The light sensor on the robot would have to be made very
> directional (large
> tube in front of it) but surely this method would call for
> only 1 light
> sensor to be used...
>
> Same for finding one's way to a 'beacon' - 2 light about a
> meter apart, the
> robot would find the midpoint (by rotating the light sensor
> constantly and
> noting the brightest points, then calculating their midpoint)
> and approach
> it, when it drives between the lights (i.e. the angle between
> them as seen
> from the light sensor is >=180) then it would know it has
> reached it's goal.
>
> Is this the case, or am I spouting rubbish?
>
> Also related to this, is how would one make the robot turn an
> exact angle?
>
> > JT
>
> Kornel Lambert (Email: kornel@black-hole.co.uk) (Mobile: 0780
> 885 3178)
> (Phone: 0181 408 0796) (PGP Key:
> www.kornel.com/pgp.asc)
> (Email-to-Pager: pager@kornel.com) (Web:
> www.kornel.com)
>
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
