From: EKF-based self-attitude estimation with DNN learning landscape information
 | Roll [\({\text{deg}}^2\)] | Pitch [\({\text{deg}}^2\)] |
---|---|---|
MLE (ours, all) | 23.139 | 18.348 |
MLE (ours, selected) | 9.657 | 4.553 |
Regression w/ L2 normalization | 25.161 | 21.996 |
Regression w/o L2 normalization | 21.668 | 18.213 |