what applicable methods can capture the motion and location(behavior) of an underwater vehicle in swimming pool?

[pooh1983]

I am searching the methods to capture the motion and location of a underwater vehicle in swimming pool.

The measured data will be used to verify the created simulation software.

The underwater vehicle can be control in six degrees of freedom.The swimming pool is about Olympic sized.

I had some ideas,however I am not sure whether they would work or not;

1. Using camera and marker,however it will only work when the vehicle is near the water surface..
2. Using camera or video cam which is located above the pool, however this technique has a problem of light refraction between water and air.
3. Using blinking signal with three receivers, this technique is good in a open water not in a swimming pool because the acoustic signal will reflect the walls. 
4. Using the pressure sensors or a dive record.

Could you suggest me to the right way? 

[johnsold]

All of the methods you mentioned could work.  And as you pointed out, each of them has its drawbacks.

It seems that most of the methods you describe only track the vehicle in two dimensions while the motion has six degrees of freedom.  If you want to track all six degrees or even the 3-D position over time a more complex system will be required.

Since the navies of the world have used primarily acoustic methods for the past 70+ years to locate objects underwater and cetaceans have been doing so for much longer, I would suggest that the acoustic approach is perhaps the best.

Put sensors at different depths as well as different horizontal positions so that the vehicle can be located in three dimensions.  In most cases the signal reflected from the walls should arrive later than the direct signal from the vehicle, so that should be relatively easy to reject.  You will not have a thermocline in the pool and the walls will be at known distances and angles to the sensors, so it should be relatively easy to discriminate between signals from the vehicle and signals from the walls.  To detect the roll, pitch, and yaw of the vehicle may be more difficult.  If the vehicle is emitting signals, put at least three transducers on the vehicle (more may be required as some sensors will be blocked from a direct path to one or more emitters) to detect those motions.

Lynn

[pooh1983]

Thank you very much Lynn for sharing your ideas. 

I still don't know how to discriminate between the further consecutive signals from the vehicle and the reflected signals from walls.

I wonder if we can encode emitting signals which can  be distinguished the order of emitting signal.

Could you advise me if you know this kind of device or techniques that can do?

I appreciate for your knowledge,information. 

[johnsold]

I think this image will help illustrate my point.  If the triangle represents the vehicle and one of the emitters is at the top vertex and the box represents the pool with a detector at the lower left corner, then the arrows describe the paths of the direct (heavy) and reflected (light) signals.  The direct path will always be shorter than the reflected path regardless of the vehicle location.  Thus, the time of arrival of of the reflected signal will be later than that of the direct signal.  So you just use the first signal from each emitter at each detector and ignore any later ones.

Simply use different pulse widths or different frequencies for the emitters to distinguish which is being detected.  You could also control the sequence in which they emit (top, bottom, right) with pauses or some kind of synchronizing signal.

Much of this will depend on how much complexity goes on the vehicle, how fast you need to get data, noise levels, and many other factors.

Lynn