mapping systems

[apitkin]

Mike's right. Traditional survey techniques are satisfactory for a basic line plot, especially when augmented by radiolocation devices like Ken Smith's 'pingers'. What we need is a device that will scan all the passage walls at a point selected by the surveyor and then save that data in internal memory for later downloading.

As it happens, there is already a device that has most of the components of your system and has the ability to receive data from external devices via infra-red, so a sonar system could in theory be added on. What's more, it goes underwater. It's called the Liquivision X1. And it doesn't have to run decompression software...

Just a thought.

Andy

[Squirrel Girl]

Barbara is Squirrel Girl on this forum.

Thanks Russell. Work is on overdrive right now, so I'm not in a position to give a coherent answer. Hah, too many sensor problems to deal with that I'm tied to with a ball and chain.

[smartbomb]

given the work i do, i have access to many manufactures of survey equipment. i have some feelers out to see if i can drum some manufacturer support. fingers crossed.

[Squirrel Girl]

They had some interesting talks at the Intl Congress of Speleology this year. Check out the proceedings at the Geophysics symposium. They had some GPS tracking of a guy in a dry cave. Then SWRI had hampster balls with sensors that they dumped into fractures, perhaps smaller than people that would float down the water.

[rchrds]

I am with mpoucher, some sort of discrete single output sensor is desperately needed to gather sidewall range information. In dry survey the disto has come into heavy use, and is very popular for this- one laser range, and with the addition of the disto-x, an azimuth and elevation to boot- all downloadable into a palm device, which plots it as you go. There is no good tool for reliably sensing ranges underwater- particularly if you are shooting down a narrow crevice to get a feel for the back- the sensors that are available are just too wide ranging, and typically give inaccurate responses (think flashlight compared to a laser.)

As for INS drift- since you can now use GPS to mitigate drift by integrating your GPS signal, which basically consists of correlating multiple signals into one discrete position, how about setting up a network of ULF emitters on the surface above the cave, which transmit on a periodic basis, and are georeferenced to whatever accuracy you need (and though accuracy of a few centimeters may be possibly between transmitter and receiver, the geolocation is typically less accurate than that.) Then collect and triangulate your position using the multiple ULF signals, and use that data in place of the GPS data to keep your drift under control.

WAY more math than I am prepared or capable of doing, but it is an idea.

Jason

P.S.- if you can come up with a single output ranging device (emitter in the front, OLED display of range on the back) in the $250 range, you will sell a whole bunch of them- I can guarontee!!

[rchrds]

Okay, brain is working today- here is an expanded example of how you could use the ULF beacons to replace GPS signals. Assuming you use a set number of beacons, and assuming that the beacons can be received off axis a certain distance (they can, but not sure how far, I haven't actually used one) and Assuming that you have georeferenced the beacons on the surface to a pretty precise position (handheld GPS will NOT suffice- either surface survey with a theodolite, or a precision GPS.)

The location is all based on the timing of the signals, similar to how GPS is calculated. ULF has a very long, low frequency wave, which makes things a little messier. But, unlike GPS, you will not be travelling fast, so updates can be in numbers of seconds, rather than miliseconds.

Assume a sinuous path in the cave. For my example, we will have three locator beacons. Each beacon is set to transmit a discreet signal- for simplicity, beacon 1 (ULF1) sends a dot-dot signal. beacon 2 (ULF2) sends a dash dot signal. Beacon 3 (ULF3) sends a dash dash signal. Each beacon signal is timed so that the terminus of the second element ends at a certain time, and the downturn of that signal is the actual time that the mobile mapper will use for calculation. Obviously, I am making an analog example that can be transferred to a digital signal, but again, not sure what the capabilities of the ULF beacons are. The signals of the ULF beacons are timed using the first (timing) block of GPS data, accurate to 10 miliseconds with good signal. The end of the transmitted signal occurs at a set known time, based on GPS time, which will be the same for all emitters, and can be continuously clocked using the GPS signal.

You set ULF1 so that its signal ends at the end of one second, ULF2 ends at 3 seconds, and ULF3 ends at 5 seconds. If the range involved is such that this is not enough time for each signal to arrive at the receiver separately, increase the times. (Not going to do the math right now.)

The whole point of making discrete timed signals, is that this allows you to have only one receiver on the mapper, as well as use only one frequency on the transmitters. Unlike a GPS, where you can squeeze 14 independent receivers on one chip thanks to high consumer demand, I'm sure you will be space constrained to be able to add multiple ULF receivers.

Now, in your processor, you build a table (or do running calculations) that calculate the distance from each beacon using the delay in the signal from each beacon, as you know exactly what time (within 10 miliseconds) that the signal was sent. In truth, having separate coded signals from each beacon is probably unnecessary, but it would help if you had to listen to them manually (analog) to determine if they were all working or something.) In addition, it allows your program to log each signal, in case some were missing.

Now that you have the distance from each beacon, you can easily triangulate your position in 3 dimensions. (creating a sphere from each beacon.) Now, of course, mathematically this provides you with, um...9 positions? but you can throw out the unlikely ones by factoring in your digital depth with some slop built in (the digital depth should be less accurate.)

I think that at this point, most of the parts required for the mapper side are available and easy to set up, as they have been used in other products. However, as ULF signal work is a pretty tightly held regimen between unauthorized users (cavers) and the FCC/US Navy/NOAA there might be a shortfall on the transmitter side- as I'm not sure what advances have been made in this field in the last few years on our (caver) side. This is something you will have to take up with the various guys that have been building them, but I doubt it is a large step, if this is your field of work.

Good luck.

jason

P.s.- also on your original block diagram, in addition to the compass you will need some measure of vertical axis- you must think 3D, not on a flat plane, if your sensor ring is canted in any direction, this will supremely effect where the results of your signals are located- the compass (horizontal axis) and elevation (vertical axis) must be integrated, and you cant expect to keep the sensor ring level with a bubble level or something- not while diving.

[rchrds]

Mike's right. Traditional survey techniques are satisfactory for a basic line plot, especially when augmented by radiolocation devices like Ken Smith's 'pingers'.

This is certainly a matter of opinion- and relates directly to the end use of the data- for paper maps used for navigation, I would agree. For volumetric calculation, for example, that is definitely not the case.

As it happens, there is already a device that has most of the components of your system and has the ability to receive data from external devices via infra-red, so a sonar system could in theory be added on. What's more, it goes underwater. It's called the Liquivision X1. And it doesn't have to run decompression software...
Andy

Andy is correct- this could very well act as a base module, providing a compass (possibly not accurate enough, and limited in vertical axis measurements, I will find out next month) a digital depth measurement, and a processor. Whether or not the processor is up to what you are going to ask of it is another discussion. Of course, you would have to integrate an additional module (ala the X-Link) that would receive and process ULF signals, add a 3rd accelerometer for measuring vertical angle, and then some way of transmitting your data from the mapper to the unit.

Probably this makes little sense to have it all separated and be forced to integrate multiple data transmission links rather than to have it all located on one board, with the calculations made in a single processor and program, and all of the data transmission hardwired.

Jason

[FW]

Hey BobK, are you reading all this?

Smartbomb, did you read the other thread about this? (High tech survey equip?) Maybe y'all should work together.

[BobK]

I've been looking at this on and off for a couple of years. I currently know how to build a system that can survey a cave with accuracy to within a foot or so. There is a slight cost issue If I could get a hundred investors, putting in around 1K each, I could build it.

Otherwise we have to wait until the electronics get better, or do the pinger thing.

[BobK]

As for INS drift- since you can now use GPS to mitigate drift by integrating your GPS signal, which basically consists of correlating multiple signals into one discrete position, how about setting up a network of ULF emitters on the surface above the cave, which transmit on a periodic basis, and are georeferenced to whatever accuracy you need (and though accuracy of a few centimeters may be possibly between transmitter and receiver, the geolocation is typically less accurate than that.) Then collect and triangulate your position using the multiple ULF signals, and use that data in place of the GPS data to keep your drift under control.

WAY more math than I am prepared or capable of doing, but it is an idea.

Jason

The INS drift in a low priced system is very very bad, on the order of meters in seconds. They are really designed to fill in the "curve" between 1 second GPS updates. A more accurate INS is orders of magnitude more expensive.

Other then INS, it might be possible to put together a system using ranging lasers if you had a highly accurate compass. This system would require two divers but could accurately map a system far better then the manual methods being used now, and would not require the divers to write stuff down.