From: EKF-based self-attitude estimation with DNN learning landscape information
 | Roll [\({\text{deg}}^2\)] | Pitch [\({\text{deg}}^2\)] |
---|---|---|
MLE (ours, all) | 31.170 | 36.967 |
MLE (ours, selected) | 9.134 | 12.932 |
Regression w/ L2 normalization | 35.358 | 45.428 |
Regression w/o L2 normalization | 36.460 | 55.347 |