Subject:
|
Re: Ultrasonic sensors - comparison & performance?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 18 Mar 2004 04:36:27 GMT
|
Viewed:
|
1094 times
|
| |
|
|
In lugnet.robotics, Brian Davis wrote:
> I've been looking at ultrasonic sensors for the RCX - two seem currently
> availible, the 24 kHz one from Mindsensors, and the 40 kHz one from HiTechnic.
> What experience do people have with these, specificly? I suspect they wouldn't
> interfere with each other, but I'm unsure of how well they will detect different
> size targets. Anybody have personal stories? What is the refresh rate on either
> sensors (3 ms? 30 ms? more?). What is the minimum size (I'm assuming angular
> here) detectable (ie - does the frequency make a difference? Is it something
> else)? How stable are the readings (variation with a fixed target at fixed
> distance? What happens when you tilt/rotate the target)?
> Anybody have advice here? Testimonials?
I cannot provide any comparison information because I think it would be improper
of me to do so.
I would like to set out a couple of technical details relating to distance
measurement using ultrasonic techniques, since I have been involved in this
technology for many years.
Range measurement.
The speed of sound is about 1100 feet per second but is dependent on air
pressure, temperature and humidity. These factors will limit absolute accuracy
buy a few percentage points. Air movement at rightangles to the beam will cause
the beam path out and back to have to be curved, increasing tha apparent
distance with increasing airspeed. (yes you can make air movement sensors using
this kind of technology :)
Measurement repetition rate.
Some surfaces, like tennis ball fluff reflect very little sound energy and can
thus be considered to be acoustically black. Other surfaces, like walls,
ceilings and other hard flat surfaces are the equivalent of an acoustic mirror
and reflect ultrasonic sound energy with amazing efficiency. The upshot of this
is that at about 1 foot per millsecond, an ultrasonic ping will bounce around
the walls of a room for a few tens of milliseconds. To avoid seeing a response
to a previous ping show up in the current ping, the ping rate must be slow
enough to permit the previous ping's reflections to have died away. This limits
the "rep rate" to something in the 10 - 50 pps range.
Beam width.
A device like an ultrasonic transmitter or receiver element has a principal axis
of emission which typical points "out the front" of the device. As you increase
the angle away from this axis, so the signal (or sensitivity) will decrease.
There is no sharp cut-off, so the general way of expressing the "beam angle" is
to define the angle at which the power has been reduced to half of the on axis
power. This is the half power beam width.
Bare in mind that a poor reflector on axis may return a lot less energy than a
good reflector quite a long way off axis.
Sensitivity.
"real sonar" system return continuous data to the host system about received
signal strength after the ping has been emitted allowing the host system to
interpet multiple returns from along the current axis path. The RCX sensor
interface does not permit this kind of sub-millisecond interface to be
practical, so it becomes necessary for the sensor to choose which echo to record
as "the echo". This then becomes a task for the sensor to threshold returns to
be sensitive enough to be able to detect acoustically "dark gray" objects on
axis while not being faked out by highly reflective objects off axis. The
typical symptoms of false off axis problems are returns from the travel surface
or features of the robot structure instead of the intended far field target.
I hope this information is of use in your review of available ultrasonic
sensors.
JB
|
|
Message has 1 Reply:
 | | Re: Ultrasonic sensors - comparison & performance?
|
| (...) How sensitive can you get this way? Specificly I'm thinking of having the sensor looking down a long "hallway" with airflow directed toward or away from the sensor (this too should result in a differing time-of-flight, yes?). (20 years ago, 6-Apr-04, to lugnet.robotics)
|
Message is in Reply To:
| | Ultrasonic sensors - comparison & performance?
|
| I've been looking at ultrasonic sensors for the RCX - two seem currently availible, the 24 kHz one from Mindsensors, and the 40 kHz one from HiTechnic. What experience do people have with these, specificly? I suspect they wouldn't interfere with (...) (21 years ago, 18-Mar-04, to lugnet.robotics)
|
9 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
|
|
|
|
