Grumpy_Mike:
I think you would be better off using lasers rather than sound. A frend of mine was once hit in a cave by a bat. Apparently they turn off their sonar off in a cave as things never change and they can remember where the walls are.
Yes, I have developed a laser based system (see the CREG Journal a couple of issues ago). The thing with laser systems is they are generally beyond me to integrate into a larger system (can't build a modulated or time-of-flight system and hacking something on to a ready build system is very hard, I've tried).
What I am looking at making at the moment is a profiler - i.e. something that can scan the passage cross section. I am pretty much there with the first part or it which is an ultrasonic/radio system where there is a base station that sends out an ultrasonic ping and the "profiler" that receives it. The idea being that the profiler sends a radio signal to the base station and starts timing. The base station receives the RF signal and sends an ultrasonic ping. The time between the profiler sending the RF and receiving the ping is proportional to distance. This means that the base station can be set up at a known survey point and the profiler moved along the passage and the distance that the profiler is from the survey point can be found.
The second part of the system is to actually scan the profile. My parameters for this are for a range around 0.6m to 5m and initially I was going to use a laser/camera triangulation approach. This has the added advantage that you also get images of the profile for drawing up maps and verification. The drawback is that the accuracy/resolution drops off with distance and with small distances between the laser and camera lens.
I too had ruled out Ultrasonic due to spurious echos etc. but then came across this array technique for narrowing the beam so it seemed worth prototyping it just to see how it went.
There are very few bats in the caves I visit (I've never see any) and certainly not in the scrotty little digs that I want to survey.