Interview question: can you design a system to safely locate and retrieve the capsule?
Part of this seems to be (sometimes?) to find radioactive sources: "The goal of this part of the exercise was to find two radioactive sources which, for training purposes, had been placed somewhere in an area of about 2,500 km2"
The Geiger counter is filled with gas and will not effectively interact with gamma ray photons; the solid state detectors interact much more strongly with them, and (particularly the semiconductor detectors) have very good energy resolution, to spot the particular energy of photons from this particular radioisotope.
Or maybe even https://georesults.com.au/product/rs-350-backpack-human-port...
I imagine they are both rather expensive.
A good intro to what it can be used for:
If you're holding it in your hand it emits enough radiation to be very unhealthy, but as soon as you put some distance between you and it then it becomes very hard to detect.
Unfortunately, Cs-137 is one of the bits of waste generated as fallout from nuclear testing, and its half-life is 30 years, so ambient levels are generally not identically zero.
Now find a road that's about 1400KM (870mi) long and pick some random point on there to drop it.
Wait a week for other passing traffic to disturb it, maybe get caught in a tire and sent even further astray.
Now go find it, remembering that even with sensitive equipment, you might need to be a few tens of meters away to pick it up out of the background radiation.
Pure horizontal scaling would be trucks with survey counters. Pure vertical would be getting the US to retask one of those radiation-sensing satellites they use to characterize stuff like the Urals trace. (Not that that'd be likely to happen - national security, etc. But have the embassy in DC call NASA anyway, just in case they do have something.) Both would be GA aircraft doing a broad survey while trucks retrace the route. In any case the logistics would be critical, and speed as ever would be costly - good thing we have government-scale resources to work with.
And, yes, the system in this case would be implemented on a human substrate, not a silicon one. This isn't Chernobyl; we don't need robots to do this job, the problem is of limited enough scope to be safely handled by properly trained and equipped humans - finding whom would be an early logistical concern. (The surveyors would be safe enough barring a forced landing or breakdown very near the source, which is unlikely but we still want to get rescue services on standby, or hire a couple of civil helicopters from the oil industry or similar, for fast recovery just in case.) And for a one-off job like this that needs to be done as fast as possible, there isn't time to design, build, and debug robots that can do it.
That's an extraordinary thing to mention in passing; do you have a source for this?
Weird assumption. Forced landings anywhere are unsafe. They are also very rare, because aircraft are quite reliable. The chances of the surveyors having a forced landing near the capsule must be minuscule (tiny danger zone compared to the extent of full search area, small probability of an aircraft having a forced landing at all).
If you are counting such minuscule dangers then you are ignoring much bigger (but still relatively small) dangers to the surveyors: they might trip during the survey, they might get into a traffic accident as they search or their aircraft might crash injuring them (irrespective of distance from the source). They might have a wildlife encounter as they are verifying a signal (false or true positive). They might suffer from heat stroke, or get sunburnt.
My intuition says that if you send out teams of people scurrying around the countryside these listed dangers are all more likely than having a forced landing right on top of the source. I also assume that you ignored them because they are all “everyday risks”, and you had tunnel vision concentrating on the radiation hazard.
I would think we need to clarify the theoretical and practical limits of the detectors.
Depending on how the sensors work the search might look very different.
If you can get some expensive lab gear and distinguish the radiation of this source from the background from hundreds of kilometers then you only need to measure at a few strategic locations to triangulate it.
If you need to get a sensor within ten meters from the source before it can be detected then you need to lug the sensor around on the ground. That would be too low for a manned aircraft to fly at.
How fast does the sensor detect? If the sensor can detect up to 100 meters but need 5 second to measure that will make the search much different than if it can measure with 500hz for the same distance.
Do we think we can sense only the source or also the places the source has been at? Does it leave a detectable trace?
What of these is hard limit by physics (nobody, for no amount of money can build a better detector) vs limit of resources (it would be very expensive to do x)? If it is a limit of resources how does detector quality scales with resources spent?
Without having solid answers to these it would be foolish to jump at designing a search method.
Fortunately for us in this case, aerial radiological surveying has long since been reduced to practice. [1] [2] The detection equipment described in those references is on the order of tens of thousand times more sensitive than consumer-grade G-M counters, and is commercially available. [3]
[1] https://www.osti.gov/servlets/purl/6084
[2] https://journals.lww.com/health-physics/Abstract/2016/05000/...
https://www.arktis-detectors.com/products/
https://www.ansto.gov.au/products/detection-and-imaging-cori...