Skip to main content
Fig. 8 | ROBOMECH Journal

Fig. 8

From: EKF-based self-attitude estimation with DNN learning landscape information

Fig. 8

Sorted samples. A number above each image is a index of a sample. In (a), top 50 samples are sorted in descending order of the error in “MLE (ours)”. Error of sample#608 in the regression model is \({ \phi _{{\rm error}}} = - 35.34\;{{\rm deg}}, \; {\theta _{{\rm error}}} = - 25.74\;{{\rm deg}}\). Error of sample#352 is \(\phi _{{{\text{error}}}} \; = \;2.80\;{\text{deg}}, \; \theta _{{{\text{error}}}} \; = \; - \;12.93\;{\text{deg}}\). In (b), top 50 samples are sorted in descending order of \(\beta\) in “MLE (ours)”. Mutual samples of the both groups are marked with red rectangles

Back to article page