Hi all,

I am trying to detect the precise arrival time of a signal.  It is exponential in nature and is
generated from an accelerometer.  It has considerable white/electrical noise on the line
already so I have 
filtered this out with a low pass 254 order FIR.  I realise this is a huge filter but it gives
a good sharp transition and the data is all post processed anyway.  
Unfortunately there is still some residual noise in the domain of the signal which means we
cannot do a simple slope-based analysis.
On the FIR output the noise is only around the noise floor level and the remaining output is
very clean thus it may be possible to do a regression / best fit to get the exponential signal
then cross correlate this to the real FIR output signal.
The signal does change in terms of its rate of rise etc but it is always exponential in shape.
Thus we can't do a correlation against a fixed shape.

Anybody know if this is likely to work or any other methods of precise arrival time detection?

Thanks,
Shane

You can post a message or access and search the archives of this group on DSPRelated.com:
http://www.dsprelated.com/groups/c28x/1.php

_____________________________________