go_away

Author Topic: Laser Rangefinding at Distances with Axon - Resolution  (Read 2205 times)

0 Members and 1 Guest are viewing this topic.

Offline ib1yysguyTopic starter

  • Jr. Member
  • **
  • Posts: 37
  • Helpful? 0
Laser Rangefinding at Distances with Axon - Resolution
« on: February 14, 2009, 02:54:48 AM »
Let me make sure I've got my math straight. I was computing the minimum distance the Axon could read a laser beamed out and returned. I'm pretty sure the math is right, but my conception is something I have questions about.

Lets start with the math.

xm / 16mhz = 299,792,458 m / s

xm / 6.25e-8s = 299,792,458 m / s

xm / 1s = 18.737 m / 1s

x = 18.737 m

So, halve that amount since it's bouncing and you can detect a wall that's 9.37 meters away with a 16 mhz clock. BUT, here's where my conceptual problem may be, so give me your input please. I'm concerned about the 'resolution' of the timer. That is to say, is that minimum distance also equal to the accuracy? Allow me to explain:

If a laser bounces at a wall exactly 9.37m away and bounces back and is read precisely 6.25e-8 seconds later, it will have accurately measured the distance from origin to wall as 9.37m. If it hits a wall 18.737 meters away, it will be read after two clock cycles and will measure the wall accurately. However, if the wall is 16 meters away it will return in 1.0674e-7 seconds -- right between clock cycles. So it wont even register until the next cycle, at which point wont it think the wall is 18.737 meters way? And won't it just continue to measure in steps of 9.37 meters since anything else returns between clock cycles?
« Last Edit: February 14, 2009, 02:59:23 AM by ib1yysguy »

Offline dunk

  • Expert Roboticist
  • Supreme Robot
  • *****
  • Posts: 1,086
  • Helpful? 21
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #1 on: February 14, 2009, 05:06:07 AM »
first of all, what you are trying to do is extremely complicated.
it may or may not be possible to build one with the tools you have available. (what tools do you have available?)

you are describing a system with a maximum resolution of 9.37 meters.
your maths looks right. the Axon's clock is accurate to far better than 1%. (you'd need to look up the datasheet on the crystal used to get an exact figure.)

if you want better resolution you will need to forget microcontrollers and build a dedicated precision timer. (ie, abandon the microcontroller for timing.)

providing you can build an accurate enough timer you still have not explained how you are going to reflect enough light back to your sensor. i'm presuming you are not going to hit a correctly angled mirror on every reading....
whatever sensor you are going to use has to be capable of responding very quickly.
avalanche photodiodes are the best choice but very expensive.
you might be able to use a more conventional photodiode for this and factor into your calculations the rise time of your sensor.

providing your sensor is accurate enough to see some light reflected back, you are going to need a way to differentiate that light from the background light.

i would recommend modulating the transmitted laser light and passing the signal from the receiver through a low pass filter to eliminate anything not at the same modulation rate.
providing you do that you will now know when a pulse is transmitted and when a pulse is received.
you can set a counter IC running when you transmit a pulse and stop it when the pulse returns.
you can now read the value on the counter IC with your microcontroller.

obviously there will be some delay before your circuit registers a returning pulse due to the rise time of your photodiode but you should be able to deal with this by taking multiple readings and applying an average value for the sensors rise time.


so while this would be perfectly possible for anyone with access to a high speed oscilloscope, PCB prototyping equipment and a lot of experience it would still take a lot of time.
i'm afraid it's one of those "if you have to ask then you probably will not succeed" type questions.


dunk.

Offline 4by4

  • Full Member
  • ***
  • Posts: 61
  • Helpful? 1
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #2 on: February 14, 2009, 12:25:25 PM »
Have you thought about trying this with sound instead of laser light, and in the process reduce your bandwidth requirements by orders of magnitude (to audio bandwidths). In general if you have never built a radar it is probably a good idea to try your hand at acoustic radar first.

I once set up an acoustic radar using my stereo system amplifier (about 70 watts per channel) and a high power tweeter. I also had a bench pulse generator, to generate about 1 millisecond pulses feeding the amplifier and tweeter. I had a microphone hooked up to a parabolic dish, feeding into an o-scope as the receiver. The dish gave it directionality. The transmission sounded like a loud click, and I could see the echoes on the o-scope and hear the echoes off of houses about 100 - 150 yards away. If 100 yards is what you want, you could do it this way. You might need ear protection as you crank up the db's.

Offline ib1yysguyTopic starter

  • Jr. Member
  • **
  • Posts: 37
  • Helpful? 0
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #3 on: February 14, 2009, 02:49:40 PM »
dunk,

I figured I was going a little over my head, but the problem is this seems like an integral part of a project I'm trying to prototype. I'm doing what I can to learn the necessary bits.

As far as a high precision timer goes, what kind of IC would one need to procure to do such a task? This needs to be small, so a laptop is out. I guess I never considered using a microcontroller to control a timer IC.

And by way of modulating laser light, I have no idea how to go about doing that. I had read about it and thought about doing something similar to FM radar, but I'm not sure if the laser component I was going to use is capable of it.

... and detecting it was something I was going to worry about later. ;) I figure I'd use some optics to amp up the signal a little and then a standard photoresistor.

Offline dunk

  • Expert Roboticist
  • Supreme Robot
  • *****
  • Posts: 1,086
  • Helpful? 21
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #4 on: February 15, 2009, 05:39:39 AM »
As far as a high precision timer goes, what kind of IC would one need to procure to do such a task? This needs to be small, so a laptop is out. I guess I never considered using a microcontroller to control a timer IC.

And by way of modulating laser light, I have no idea how to go about doing that. I had read about it and thought about doing something similar to FM radar, but I'm not sure if the laser component I was going to use is capable of it.

using a laptop would be a step backwards.
while the CPU might run at GHz speed the I/O ports do not. you would still have to build the same circuitry to interface your project to it.

start reading up on counter ICs. google "counter ic"  brings up a few good links. for example:
http://www.doctronics.co.uk/4510.htm
you could drive such a counter with a crystal allowing you to accurately time outgoing and incoming pulses.
i do not know if counters exist which run at the sort of speed you will need but you have to start reading somewhere.

i suspect this part of your project will take you several years to complete so you would be well advised to start learning about all aspects of electronics now.
as 4by4 says, try making a sound based prototype.
you will need the knowledge if you are to have any hope of building circuitry that can operate at the speeds you are looking at.

Quote
... and detecting it was something I was going to worry about later. Wink I figure I'd use some optics to amp up the signal a little and then a standard photoresistor.

i suspect finding a sensor within budget will be the hardest part of your project. the reflected light from a laser is minimal compared to the lighting in a typical indoor environment.
read the datasheets on standard photoresistors. they have nothing like the response time you are looking for.


despite my general pessimism on the matter, there are a few success stories out there:
http://innovexpo.itee.uq.edu.au/2001/projects/s348491/thesis.pdf

have you considered using a different method than time of flight?
here are some notes on my geometry based laser ranging efforts: http://mrdunk.googlepages.com/sensors
i have not worked on that project for a while... i must get back to it one of these days.


dunk.

Offline 4by4

  • Full Member
  • ***
  • Posts: 61
  • Helpful? 1
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #5 on: February 15, 2009, 08:02:41 AM »
Dunk, that Australian paper is interesting, if a bit sparse on details. Also, notice that the gentleman wasn't able to get it working, at least that was my impression on a quick scan. And he probably had lots of help from his professor.

About counters - there are very fast counter ICs. They can count cycles of RF signals directly. There are some simple products that make use of these. I think the front end of these is just a pre-amp connected to a counter.

http://www.supercircuits.com/Wireless-Devices/RF-6
http://www.thespyshop.ws/ns-15.html

One thing to keep in mind about circuitry operating in the hundreds or thousands of MHz is that you can't use ordinary circuit boards. The circuit traces behave as transmission lines at those frequencies, so you have to understand at least some basics of transmission line theory. The design technique for RF ciruit boards is called "microstrip" design. It's a difficult subject.


Offline Soeren

  • Supreme Robot
  • *****
  • Posts: 4,672
  • Helpful? 227
  • Mind Reading: 0.0
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #6 on: February 15, 2009, 05:39:57 PM »
Hi,

About counters - there are very fast counter ICs. They can count cycles of RF signals directly. There are some simple products that make use of these. I think the front end of these is just a pre-amp connected to a counter.

http://www.supercircuits.com/Wireless-Devices/RF-6
http://www.thespyshop.ws/ns-15.html

The bug detectors has got no counters. They're not much more than a resonance circuit (with bells and whistles of course) and an amplitude meter.
Regards,
Søren

A rather fast and fairly heavy robot with quite large wheels needs what? A lot of power?
Please remember...
Engineering is based on numbers - not adjectives

Offline ib1yysguyTopic starter

  • Jr. Member
  • **
  • Posts: 37
  • Helpful? 0
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #7 on: February 16, 2009, 02:18:02 AM »
Isn't the easiest way of measuring distance just doing calculations with the phase shift of reflected light? If you know the modulated wavelength of the light emitted, the light returned will be offset a bit. The question is how do you detect just the laser light through a beamsplitter? I don't eve know what kind of sensors you'd use.

An equasion like this should work to determine the time of flight

MeasuredPhaseShift = 2pi x TOF x ModulatedFrequency

=

MeasuredPhaseShift/2pi x ModulatedFrequency = TOF

Once you have TOF you can calculate distance. But how do you do this mechanically? Is there some kind of sensor I could get to make it easy?
« Last Edit: February 16, 2009, 02:24:01 AM by ib1yysguy »

Offline ib1yysguyTopic starter

  • Jr. Member
  • **
  • Posts: 37
  • Helpful? 0
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #8 on: February 16, 2009, 05:09:06 AM »
Okay okay. Lets do the easy thing and consider this:

What if we just measure parallax? Say you wanted to make a spotting scope style range finder. You could make the optical tube with a beamsplitter and a laser inside around 830 nm so you can't see it.

Use a beamsplitter to reflect the optically magnified image through a bandpass filter (http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productid=1903&showall) leading to one of these
http://www.sparkfun.com/commerce/product_info.php?products_id=8668

So the camera will always see the field of view you're looking at through the optical tube, regardless of zoom. The image will be divided into a grid of 1300x1040 pixels. The only light it's going to see is the bandwidth of 830 nm projected by the laser (which you wont see). Using that grid, you could simply calculate the distance using some trig, knowing the distance of the laser to the beamsplitter because the grid will show the dot of the laser at varying distances from center depending on distance.

Easy, right?

Offline 4by4

  • Full Member
  • ***
  • Posts: 61
  • Helpful? 1
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #9 on: February 16, 2009, 09:19:05 AM »
Let me rephrase - the radio frequency counters do have counters otherwise they couldn't measure frequency to n digits. A resonant circuit can't do that. Those examples I linked to may not be true radio frequency counters, they may be imitations, but they do exist. Amateur radio guys use them a lot.
« Last Edit: February 16, 2009, 09:31:56 AM by 4by4 »

Offline dunk

  • Expert Roboticist
  • Supreme Robot
  • *****
  • Posts: 1,086
  • Helpful? 21
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #10 on: February 16, 2009, 11:31:46 AM »
Okay okay. Lets do the easy thing and consider this:

What if we just measure parallax? Say you wanted to make a spotting scope style range finder. You could make the optical tube with a beamsplitter and a laser inside around 830 nm so you can't see it.

Use a beamsplitter to reflect the optically magnified image through a bandpass filter (http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productid=1903&showall) leading to one of these
http://www.sparkfun.com/commerce/product_info.php?products_id=8668

So the camera will always see the field of view you're looking at through the optical tube, regardless of zoom. The image will be divided into a grid of 1300x1040 pixels. The only light it's going to see is the bandwidth of 830 nm projected by the laser (which you wont see). Using that grid, you could simply calculate the distance using some trig, knowing the distance of the laser to the beamsplitter because the grid will show the dot of the laser at varying distances from center depending on distance.

Easy, right?

yup. you have a far better chance of success using this method.
avoid using non visible lasers if you can. the human eye has no blink reflex at 830 nm so it would be easy to damage peoples eyes without them ever knowing how.

even though you use a bandpass filter you will still get some background light at the same wavelength as your laser so some modulation would help against false positives.

but yes, the only tricky thing about this method is interfacing the CMOS camera. after that it becomes a software problem.

if you want to cut down on your development time look into the http://www.surveyor.com/blackfin/ camera.
it has already done much of the hard work for you. i think it even has I/O pins to drive the laser.


dunk.

Offline chelmi

  • Supreme Robot
  • *****
  • Posts: 496
  • Helpful? 15
Re: Laser Rangefinding at Distances with Axon - Resolution
« Reply #11 on: February 16, 2009, 10:58:32 PM »
http://www.dealextreme.com/details.dx/sku.1140
16USD, free shipping.

Probably crappy quality but for the price it is worth try it :p

 


Get Your Ad Here

data_list