electronic distance measure

[gregcarter]

We have a rolling platform that rolls around on a large concrete floor with no obstructions. We want to know where it is say 1 to 5 times per second for steering controls.
Were trying to keep it rolling in a straight line with very little deviation.
We have considered robotic geodimeters, however their coordinate reading ability is advertised at over 5 seconds, and they are extremely expensive.
It seems that 2 transmitters could be set up at a known distance apart, probably at a different frequencies to tell a 2 receivers on the platform what the distance is from them, and then calc coords for comparison with theoretical locations, and then steer accordingly.
The geodimeter company does advertise a tracking rate of .4 seconds, which may be fast enough, but the data still has to be sent to the platform and processed.
The area the platform travels is anywhere from 6 feet to 1000 ft. We can always cut down the length of travel if the accuracy goes away.
Any help on this would be appreciated.
Thanx ahead,
Greg Carter

 

[BrianG]

From your description of the requirement of "travelling in a straight line" and no obstrcutions on the floor, have you considered a laser based system where the platform can follow a placed laser line? If the platform must roam independently and prior placement of a laser marker is not possible then perhaps the platform could use a bar code type system reading graduations on the floor?

If you are trying to measure position by relative signal strength between a pair of transmitters, one of the problems with using r.f., even at microwave frequencies, is trying to get consistent propagation and relative signal strength measurement.

Although there are ways to use phase encoding to get quite good resolution, neither this nor a radar type approach would not seem to fit your requirment of 6 feet as the smallest distance required.

 

[Comcokid]

To comment on BrianG laser idea - I saw in the late 80's a system that was used on 1D linear cranes based on a HP surveying-type EDM laser with a mirror on the crane. The industrial process instrument lasers may not have the range but the surveying lasers do - and at resolutions below 1mm. In this sytem the laser closed the loop in the motor control system to allow heavy loads at the end of the crane to stop-on-a-dime without swinging.

Zigbee is promising the ability to do loction determination within a building down to +/- a few feet. Freescale has demonstrated a ZigBee/UWB transceiver combination that should be much more accurate when it becomes available.

 

[VisiGoth]

The previous suggestions about not using RF is good advice. Precision direction finding will have a phase interferometer component as the amplitude based systems are just good for ough direction (in practice). It takes bandwidth to get good time estimation (which is phase, which can be turned into position and direction). To operate in a licensed band would (IMO) require the use of an existing protocol, as Comcokid gave an example of. To roll your own could be fraught with difficulties like multipath. So, I would concur with the previous authors that you would do well to look elsewhere, unless there is a practical system in place that you can buy and not take on any development risk.

 

[zappedagain]

"anywhere from 6 feet to 1000 ft." - what accuracy do you need? +/-1 ft out of 6? 10ft out of 1000? Is that in the travel direction? What about the in the direction perpendicular to travel ("very little deviation" - we need numbers!)?

Also, how fast do you plan to travel; how much area will you travel between measurements (1 to 5 per sec)?

Laser methods can be extremely accurate, maybe more than you need.

Can you follow a wire path cut into the floor? Robots have done that for years. Or how about traversing a grid of wires in the floor?

Or how about sensing the rotation of each wheel and calculating your direction from that? Maybe some reference points along the walls can recalibrate your movement every so often.

There might be a lot of mechanical or electromechanical solutions at lower cost. Check out a Roomba from iRobot.

 

[gregcarter]

Really, really appreciate your help on this project.
The square platform has 4 wheels in 1 direction and 4 wheels in the other and prints lines on the floor, usually rectangles and parallelograms with other rectangles inside them and lines crossing horizontally and sometimes vertically.
90 degree +- direction is changed by raising one set of wheels.
Small angular correction steering is done by skid steering stepper motors on each side of the platform. 4 motors total. 2 motors drive and steer in one direction, while the other 2 rest until the platform stops, 1 set of wheels is lowered, and the platform is driven at approximately 90 degrees from the original direction.
I erred with the 6 foot requirement. The transmitters or geodimeter could be placed anywhere. The smallest rectangle is around 6x10 feet, the largest is usually 28 by 40 ft. However when we place many of the largest rectangles together, the entire pattern can be 1000 ft long by 40 or so feet.
If the RF idea isn't practical, I'll have to work with the geodimeter.
We are using a geodimeter on a concrete leveling machine (laser screed) that uses a plot of the finished product. The geodimter tells the onboard computer where it is and the leveling machine adjusts one end of itself to the required grade, the other end follows the level of the last pull.
This machine is used to level concrete on large sloping areas. It only needs one piece of information for the plot about every 60 seconds, so it's not too busy giving positioning info, but the robotic head follows the screeds every move.
I've considered the Trimble 5600 series geodimeter for my device, but their specs indicate that it may be too slow to relay position. I really need to call them for information on their product.
I checked out the Zigbee and Freescale websites. What an eyeopener.
Thank you again for your advice,
Greg Carter

 

[gregcarter]

I forgot to mention, although the speed isn't critical, fast is best (4 to 6 mph.), but average walking pace would be plenty good (2 mph), and even 1 mph would work.

I was also incorrect about the sloping laser screed, as the geodimeter continually sends location information to the onboard computer that then tells the machine the correct grade to adjust itself to.

Here's a link to see what it looks like:

http://www.somero.com/products/3-dProfiler.htm