Simulation of creating EFFICIENT fingerprint database for UWB Indoor Localization

Hi all,

Is it possible to create a small program (in Matlab, preferably)-- thinking about Monte Carlo simulation--that creates a fingerprint database?
I would probably want to record/store 2D or 3D location information along with Channel Impulse Response (CIR) at that point.
Now, I am not clear what precisely of the CIR could/should I try to record? Is it power delay profile? Or is the CIR itself directly measurable? Or.. something else?

The CIR database creation only half of the story.

I then want to find a quantifiable item in that database and compare it with standard "item" that has been theoretically proposed. Could it be Cramer-Rao Lower Bound (CRLB)? Something else that can be picked up easily?
How do I extract that item or CRLB, for example, from my simulation?

I am so confused, but I must get it done. My goal: simulate an efficient (=low complexity) CIR fingerprint database creation using some kind of algorithms.

Gurus, old hands in this area, including Matlab/simulink simulation creators, please show me a way. Right now, even some keywords help. A roadmap would be great!

Thank you for your time and input.

Cheers,
tienshan

I think I need to narrow down what I want to do.
Today I thought may be I should simulate power delay profile as a fingerprint for my indoor localization project, using a ray tracer.

I found a ray tracing tool written in Java, called RAPSOR, in Sourceforge.
Unfortunately, I could not make it work (it needs netnbeans) and there is no documentation for a beginner like me.
Has anybody used this tool before?