interpolating large amounts of data

[systems_eng]

Does anyone have any recommendations regarding a method for interpolating large portions of data.  From what I've seen through my testing, is that I'm able to sample a lot of points (incoming messages from a gps device).  My problem is when I go to interpolate these sets of data.

I've tried capturing a (lat, long, timestamp), then interpolating both lat/long togetherand one at a time.  I've also tried storing the readings to different .xls files, reading the data back into LabVIEW and then interpolating together, and separately.  But niether of these methods have been able to avoid the lack of system resources issue, when the program has captured near 4000 signals.

I'm not sure where else to go from here.

All these methods have

[Wiebe@CARYA]

"systems_eng" <x@no.email> wrote in message news:1161119411978-428898@exchange.ni.com...
Does anyone have any recommendations&nbsp;regarding a&nbsp;method for&nbsp;interpolating large portions of data.&nbsp; From what I've seen through my testing, is that I'm able to sample a lot of points (incoming messages from a gps device).&nbsp; My problem is when I go to interpolate these sets of data.
&nbsp;
I've tried capturing a (lat, long, timestamp), then interpolating&nbsp;both lat/long&nbsp;togetherand one at a time.&nbsp; I've also tried storing the readings to different .xls files, reading the data back into LabVIEW&nbsp;and then interpolating together, and separately.&nbsp; But niether of these methods have been able to avoid the lack of system resources issue, when the program has captured near 4000 signals.
&nbsp;
I'm not sure where else to go from here.
&nbsp;
All these methods have



4000 points is not that much. Interpolation (linear?) should take about 1 sec. max. What lack of system resources do you get? Working memory, memory leaks or proccesor load?


I'd be hard to recommend a solution without seeing some code.


If you'd be able to post 1 vi that has the input data as default values, and a working solution with leaks, we can help you much better. (In LabVIEW 7.1 for me, but someone will help you if you post in LV8, or 8.2)


Are you using real excel files, or are the .xls files tab dilimited asci files?


Regards,


Wiebe.

[systems_eng]

Okay, so here's some of the code I'm looking at.  I don't know that the actual acquisition of the gps data is important, but I included it in there just in case.  You'll see that I'm using a cubic hermite interpolation on the data, and a ramp function on the time.  You mentioned that 4000 samples wasn't that many, but I do have to resample the data at .001 accuracy as it ramps up.  So that's another 1000 samples in between all of the captured gps data.  I noted in the program that program is rejecting the ramp function when the sampled data reaches large portions.

I don't know that some of the files converted cleanly, so I'm posting in LV8

I believe I'm using .xls files tab dilimited asci files.

The only other thing to consider is whether or not this is the best method, for interpolating such data.