Skip to main content

Table 5 Trajectory RMSE(m)/backend processing time per frame(ms) on the Kaist Urban dataset

From: M2C-GVIO: motion manifold constraint aided GNSS-visual-inertial odometry for ground vehicles

 

Kaist 30

Kaist 31

Kaist 32

Kaist 33

Kaist 38

VINS-Mono

201.44/20.02

343.69/14.18

34.72/11.29

23.07/21.29

88.00/23.90

Pure GNSS

10.53/\(\times\)

7.47/\(\times\)

4.85/\(\times\)

9.01/\(\times\)

15.26/\(\times\)

GNSS-MSCKF

9.88/4.65

25.44/3.90

8.33/6.18

10.88/5.21

10.41/5.12

VINS-Fusion

39.63/27.52

23.47/18.17

6.79/14.65

5.25/27.6

10.28/27.93

Ours

6.92/5.01

4.46/4.73

2.65/6.59

3.80/5.66

7.89/6.34

  1. Bold values mean the best one of RMSE for each sequence