Subject:
|
The IR beacon, was Re: Homebrew sensors?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 13 Nov 1998 22:32:55 GMT
|
Original-From:
|
Paul Haas <paulh@hamjudo.com/antispam/>
|
Viewed:
|
2881 times
|
| |
| |
On Fri, 13 Nov 1998, Daniel Miller wrote:
> Or, if you wanted something that looked cooler... the sound sensor is on a
> turret that rotates as above (with an angle sensor keeping score). When
> it finds and returns to the max signal, the robot would then turn under it
> while the turret remained stationary. You'd have to limit the motion of
> the turret from +180 to -180 so as to avoid winding up your wires.
> (You could test the program for either of these using the light sensor
> instead... build a tunnel around it and it becomes directional.)
The light sensor is directional without a tunnel. I use the IR tower as a
beacon. I've got a perl script that listens for RCX messages. Here's the
normal sequence of events:
1. RCX sends the message 2 "Turn on your beacon, so I can find you"
2. PC responds with RCX message 3 "Ok, I will turn on the beacon"
(Eventually the RCX will try again, but that code isn't in there
yet.)
3a. PC pauses for .25 seconds, the IR tower is dark except the green
LED
3b. RCX takes a baseline reading (call it Baseline)
4a. PC sends out 240 characters of 0x00, (1 second at 2400 baud)
4b. RCX takes 3 readings at .2 second intervals.
it keeps the highest reading (call it Max)
it subtracts the baseline reading (call it Delta)
Steady light sources don't cause problems
5. PC sends the message 6 "You can send RCX messages again, I've
turned off the beacon"
(There is a pause between beacon mode and RCX messages. It's
probably not needed. I plan on sending a couple 0x55 bytes to
get the levels right and drop the delay.)
6. RCX sends the measurement back to the PC has an RCX value (for
debugging)
7. Paul reaches over and moves the tower or the RCX, because he
hasn't written the rest of the program. Go to 1.
This works fine in my living room. The data is good. Even at 8 feet, the
IR tower always gives the highest Delta. Some reflections are fairly
strong, so absolute values aren't usefull. I need to tune the timing.
More than a second per reading is too much.
I think that the light sensor is read once every 3 milliseconds. When you
think you're reading the light sensor, you're really getting the last
sample. Anyone want to verify that?
I have to read the light sensor multiple times because sometimes I sample
during a stop bit. The light is on for zero's, off for ones and stop
bits. (maybe the parity bit too, not sure).
My robot is the classic 2 motors (1 per wheel), and a bump sensor. It
works fine running Torbot programs. I stuck a light sensor on it.
I plan on making a "garage" for the RCX. An otherwise dark place with the
IR tower. The RCX will know it is "home" when the IR beacon is very
bright and it is otherwise dark. I get beacond on readings of 70 to 100
at 6 inches depending on alignment.
--
paulh@hamjudo.com http://www.hamjudo.com
The April 97 WebSight magazine describes me as "(presumably) normal".
|
|
Message has 2 Replies: | | Re: The IR beacon, was Re: Homebrew sensors?
|
| (...) ph> I think that the light sensor is read once every 3 milliseconds. When you ph> think you're reading the light sensor, you're really getting the last ph> sample. Anyone want to verify that? Dunno about the bit level, but the Official Lego (...) (26 years ago, 16-Nov-98, to lugnet.robotics)
|
Message is in Reply To:
| | Homebrew sensors?
|
| I've been brainstorming for a while about what could be done with a kludge. I'm more into the whoosh-bang side of engineering than the beep-zap, so I don't know if it would work... A few years ago, Estes (the model rocket people) made a "rocket (...) (26 years ago, 13-Nov-98, to lugnet.robotics)
|
13 Messages in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
Active threads in Robotics
|
|
|
|