The Challenge is to locate a potentially injured Coal Miner, in a Coal Mine that has just collapsed, using some kind of technology (You don't want to send in even more people into a dangerous location, if you don't have to do so).
Sounds easy doesn't it?
The first answers is always "Use GPS". Satellites orbiting the Earth do not penetrate under ground far enough to be of any value. So GPS approaches are dead at the Starting-Line.
Next problem is that Coal absorbs most radio frequencies.
The physical environment also has several challenges:
Dust so thick that you can not see through it.
Work shifts last twelve hours, which makes powering anything a challenge, and you are not allowed to replace the batteries in most areas of the Coal Mine. Also there are extremely stringent regulations on the use of Lithium Batteries; see the MSHA FAQ in the menu.
There are few to none, power outlets where you could plug anything in to power it.
There also issues of Intrinsic Safety, where available energy is extremely limited. Your total energy budge must not be capable of igniting a Methane Gas atmosphere.
Finally it must be small and lite physically, would you want to carry your design around for twelve hours?
Keep in mind that idea is to location someone after the mine has suffered some type of catastrophic damage. This means that placing 'way points', or anything like that is not the way to go for reference points, as there is no way to know how they might be damaged. This doesn't rule reference points out, but you can not rely on them. Running wires (Leaky Feeders for example), or Fiber, of any type is also not going to work, as they are prone to being damaged.
Do you have an idea that no one has yet thought of?
Someone once asked me if "you had had six to seven million dollars to spend on for research contracts in trapped miner communications and miner tracking systems for miners, what would you do with it?".
These were my suggestions, as to how I would use the money:
Rogers work was used in World War One to communicate with submarines. It works well underground. There is also similar work by a contemporary fellow named Minto.
Rogers work is known by a few other names and related research such as the Aharanov-Bohm Effect (Physics), Poynting Vector (Mathematics), Scalar Waves (Pseudoscience). The differences comes down to issues of geometry but I've not got my head wrapped around all of that yet.
I can't put my finger on the exact citation at the moment but T.E.Bearden made the comment that the J.H.Rogers system is rediscovered every few years then "Lost" (it worked to well). Classic RF energy falls off at 1/(r*r) where "Scalar Technology" falls off at 1/r. Most of this is still considered in the Pseudoscience category, but the main stream is catching up as they always do a decade or two later.
I would sponsor a Circuit Cellar contest. These contests attract the "best and brightest" of the Engineering World, from around the world. Someone in this group might have a idea that no one in the industry has thought of, because of their backgrounds in different fields.
A Circuit Cellar contest, requires a $150,00 sponsor ship fee to handle all of the administration, tasks such as promotion and judging.
Offer a prize fee of $250,000, the highest ever offered, to the wining entry that can come up with a workable design for finding a trapped Coal Miner. There would also be a few smaller prizes. The way I see it for, $500,000 you get thousands of people around the world competing to solve, what for the moment, is an unsolved problem. With that prize money as motivation, the solution will probably be quick!
RFID-RADAR offers a system that can do location in 3D, based on Angle of Arrival.
Tests with WWVB at 60 kHz to see how deep low frequencies penetrate, and what is the highest frequency that will give the greats usable depth.
Synchronous Demodulators in relation to Sutton & Spaniol et.al.'s Black Hole Antenna. Dr. Sutton, retired from NASA, has given me permission to post some of our email exchanges. In-a-nut-shell this technique lets you pull signals berried in noise, without using large inductors (coils).