One form of "smoothing" you could do is a "moving N-point Average", a form of low-pass filter (with a little Signal Theory, you can easily work out the filter characteristics ...). What you do is have a circular buffer (do you know what this is? an N-point Array where you add new elements in "circularly", replacing the oldest) and report the average of the N points.
Here's an illustration of how it works. Suppose you had a switch that turned a light on and off (so your measured values were 0 or 1). You make measurements once a second and want to know the average illumination. Well, any single measurement will be 0 or 1, and you can say "to the nearest second" whether the light is on or off. Now, suppose I turn it on and off at random intervals, roughly once/second. I'll still measure 0 or 1, but now if I create a 1-minute circular buffer, after a minute, I'll have 60 measurements, and probably about half will be 0, half will be 1, so the average illumination will be 0.5. The next second, a new measurement replaces the oldest (either keeping the average the same or changing it by 1/60) so my "estimate" of the illumination will change slowly (and will represent the illumination for the past minute -- low-pass filters look "backwards" in time). In particular, if the bulb burns out, the reading won't go to zero until about a minute later.
This form of moving average is simple to implement. There may even be VIs supplied by NI that do it for you (I tend to "roll my own"). Just remember there is a trade-off between "smoothness" and "responsiveness".
Bob Schor
