Table 3 Reduplicate experiments results of random noise within the Zara dataset.

From: Goal-oriented autonomous decision-making for social robots via collaborative interactive inverse reinforcement learning approach

Method

Completion time (s)

Success rate (%)

Collision rate

person (%)

Collision rate

Wall (%)

GO-ADM(0.01)

27.3 ± 1.1

90 ± 3

10 ± 3

0

GO-ADM(0.05)

27.4 ± 0.9

88 ± 3

12 ± 3

0

GO-ADM(0.1)

27.8 ± 1.2

86 ± 4

14 ± 4

0

GO-ADM(0.2)

28.4 ± 1.1

85 ± 3

14 ± 3

1 ± 1

GO-ADM(0.5)

28.8 ± 1.5

82 ± 7

17 ± 6

1 ± 1