Table 2 Relative errors [%] using GeodesyNets and a comparable mascon representation.

From: Geodesy of irregular small bodies via neural density fields

  

Bennu

Churyumov–Gerasimenko

Eros

Itokawa

GeodesyNet

low

0.72

2.30

1.82

2.13

 

hi

0.02

1.75

0.17

0.38

masconCUBE

low

1.00

2.87

2.39

2.47

 

hi

0.01

2.03

0.13

0.70

  1. Close to the surface, the advantage offered by the use of GeodesyNet is revealed (bold numbers indicate the best model).
Back to article page