pub fn smooth_with(
log: &FlightLog,
readings: &Readings,
options: Options,
) -> Result<Smoothed, Withheld>Expand description
Smooths a flight log whose readings are already taken.
The climb’s port coefficient isn’t learned (PORT_PRIOR_SD): the log is smoothed at its
prior mean and again a standard deviation off, and each standard error σ becomes
hypot(σ, x_off − x), the smoother’s linear response to the coefficient’s spread (a consider
analysis: S. F. Schmidt, 1966; D. Simon, Optimal State Estimation, Wiley, 2006, §7.6). The
second run replays the first’s choices, so the difference is the coefficient’s effect alone.