Table 1 Average RMSEs in training set (\(N=252\)) for each channel with 95% CI and p-value
From: A physics-based neural network for flight dynamics modelling and simulation
FlyNet | Classical | p -value | |
|---|---|---|---|
u [ft\(s^{-1}\)] | 2.222 [0.327, 6.392] | 3.623 [0.753, 22.473] | 0.000 |
w [ft\(s^{-1}\)] | 1.288 [0.339, 3.723] | 1.626 [0.455, 8.903] | 0.000 |
q [rad\(s^{-1}\)] | 0.012 [0.004, 0.029] | 0.018 [0.005, 0.087] | 0.000 |
\(\theta \) [rad] | 0.023 [0.006, 0.064] | 0.038 [0.009, 0.213] | 0.000 |
v [ft\(s^{-1}\)] | 2.159 [0.502, 6.169] | 3.112 [0.653, 14.319] | 0.000 |
p [rad\(s^{-1}\)] | 0.024 [0.006, 0.057] | 0.035 [0.010, 0.200] | 0.000 |
r [rad\(s^{-1}\)] | 0.021 [0.006, 0.053] | 0.032 [0.008, 0.136] | 0.000 |
\(\phi \) [rad] | 0.044 [0.009, 0.138] | 0.074[0.021, 0.648] | 0.000 |