After a fresh round of ops testing, using the calibration for water speed here, I've found that the calculated VDR (Tidal Set and Drift) is much, much more accurate than before. However, there is still a minor issue: when I set out to test it, it was slack tide. Meaning, the tidal speed (drift) was close to zero, and the course (set) was nearly impossible to verify. Most programs only calculate two directions--max flood course, and max ebb course. At slack tide, when the current is shifting, the direction is going to be somewhere in between the two. So it was a little difficult to verify the course information, but one thing became clear--the need for some sort of data normalizing, or damping, to average out and minimize the noise. I have yet to develop this, but I imagine it will involve either a change of heading or rate of turn. In any case, expect a big update soon to the overall code. And here are the results from the tests, first in a table form and then the screenshots. Also, the code was updated to refresh True Heading at 10hz, so as to be compatible with certain radars.
|Predicted Set||Predicted Drift||Calculated Set||Calculated Drift||Set Error||Drift Error||Drift Error %|