Using polarization for relative attitude determination

As it may be obvious from my question, I am novice to optics. And I'm aking friendly expert advise. My hypotethical scenario is sketched to learn about how to use polarized light in indoor and outdoor robotics applications.

What I would like to do is to determine the relative attitude of two mobile robotic platforms, or the attitude of a mobile robot relative to a stationary node. It is easier to explain my question by using the second example

Imagine and that we have a light source projecting vertically polarised light to a wall or to a screen. Assume also that we have a receiver sensor on the mobile platform with a polarising filter in front of it, and it is pointed toward to the same wall/screen. Depending on the relative angle of the polarised light source the receiver's polarizing filter, we will measure different intensities.

If we rotate the receiver's polarising filter continuously, in ideal situation, I would expect to observe a sinusoidal wave form for the measured light intensity. Instead of dealing with the absolute intensity values from the receiver sensor, if you detect the peak of the sinusoidal wave then we may find the best alignment with the projected light and the receiver polarizations. Since we know the rolative angle of the receiver filter with respect to the our mobile platform, we can find the relative angle/attitude of our robot with repect to the light projecting node.

Now if we assume that the polarised light projector is mounted on a mobile robot and it is pointing to another robot, then we can find the relative angle of these two robots.

Questions; 1. If I put a linear polarizing filter (removed from low a cost 3D glasses, any suggestion from where I can get a better one -low cost please-) infront of our light source and project the light to an ordinary wall, due to the uneven surface of the wall will we lost the the polarization ? 2. How to make a wall/screen which preserves the polarization of the projected light? 3. Can we use either ordinary visible light (a light bulb) or a bunch of IR LED as the light source behind the polarizing filter? (Do IR light from the LED can be polarized as normal light?) 4. In an outdoor situation, if we have a long flat road and two mobile robots on opposite ends of the road, pointing to each other, can we use the above mentioned method to detect the relative angles between the robots? 5. How daylight effect our polarized light experiment? 6. Can you suggest an "optical" method which allows us to determine the rolative angle of two mobile robots seperated from each other by

100-200 meters? (Yes, it sound like a big ask!) 7. Thanks in advance for the friendly advices.

Cheers,

Miem Chan

Reply to
Miem
Loading thread data ...

Why does it have to be optical? Sounds like an ideal application of low-cost accelerometers, unless there are special circumstances you didn't describe.

You can use a laser to communicate between the two robots, if optics in the loop is a "must" :)

To do it your way, I would suggest an added complication - modulate the light source (say, 10kHz) and filter the other end to allow through only this modulation before you start looking for those sinusoidal peaks.

I take it you are assuming an acute angle here.

Reply to
zwsdotcom

Unfortunately, this relative angle is coaxial with the direction of beam propagation; thus it isn't useful in most mobile robotics tasks.

Check out edsci.com or scientificsonline.com for polarization filters. Depending on the surface properties, polarization may or may not be lost when reflecting off your rough wall. In fact, light can be polarized by reflecting it off some materials at a shallow angle; hence "Rayban" sunglasses have polarized lenses to help reduce this glare. At steep angles (e.g. nearly perpendicular), I believe most surfaces preserve polarization.

Check readily available surfaces with a pair of polarization filters until you find one which fits your needs. Reflecting light off a screen will greatly increase the power requirements.

All light can be polarized. Polarizing filters block a certain percentage of the light passing through them. Lasers can be designed to generate polarized light, but these will generally be expensive.

It will flood your optical sensors with stray light. You may need to modulate your signal to help separate it from background noise.

The dominant problem you will face is simply tracking. At that distance, you will want to use lasers pointed directly between the robots; other light sources would require hundreds of watts to operate. Once you have a tracking system in place, you could query it for the angle it is targeting. Use polarized light or two parallel beams to obtain the rotation about the beam axis.

- Daniel

Reply to
D Herring

And somewhat hard to find now on the surplus market. In the old days, laser printers and some laser disk players used polarized he-ne lasers (using a Brewster window rather than a filter), and if you can locate one in the junk pile you might find the tube still works. A common polarized tube was the Toshiba LG-3217. Which reminds me that I've been meaning to unload my cache of laser tubes and power supplies, which include the 3217...anyone hit me up via EMAIL (not this list) if you're interested.

A good hacking page for those into the good ol' days of glass laser tubes:

formatting link

-- Gordon

Reply to
Gordon McComb

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.