One thing you might be looking for is consistency. For instance, when calibrating a radar system, you "tune" it until you get a clean signal. You can then further "tweak" multiple stations to ensure that they give matching values.
In the 1701-D, there are very few unique individual sensors... rather, you have lots of copies of any particular sensor type. So, say you have 50 "tachyon detection cells" in various places on the ship. You tune them all to the same frequency (or whatever you do with a tachyon detector cell) and focus them on a single target. Ideally, all should give the same resultant data. In reality, there will be inevitable variation between the return data of each cell. You'd determine the mean value, ideally compare it to the "known true value" (which isn't always impossible if you're in an unknown place... for instance, if doing spectroscopy, you can be confident that the color lines for hydrogen, say, are always going to be in the same place!) and then adjust the various devices to agree with each other and whatever "known" values you have as reference.
It really wouldn't be very computationally-intensive, it seems to me. It would, however, make the ship largely blind (or at least blind with a particular type of sensor) for a period of time.