Table 2 Generalization performance evaluation in Test Scenario-5

From: Cognitive embodied learning for anomaly active target tracking

Method

30% occlusion

40% occlusion

50% occlusion

60% occlusion

70% occlusion

 

SR(%)

RPL

SR(%)

RPL

SR(%)

RPL

SR(%)

RPL

SR(%)

RPL

NBV

2.8

105.7

1.9

117.5

-

-

-

-

-

-

EI (STF)

9.7

56.3

2.7

97.4

-

-

-

-

-

-

EI (SI)

18.9

21.7

10.6

49.6

3.8

108.3

-

-

-

-

EI (EA)

66.8

15.1

60.9

18.9

52.7

21.8

45.4

24.6

38.0

28.3

EI (EA+SI)

76.8

14.9

70.2

17.3

63.6

19.9

58.4

22.4

50.1

25.1

CEL (Rules)

90.0

11.9

86.0

12.1

82.3

12.6

79.2

13.1

76.8

13.7

  1. The values in bold indicate the best-performing results in every setting. SR is the success rate, and RPL is the relative path length. EL(EA + SI), EL(EA), EL(SI), and EL(STF) represent four common embodied learning (EL) paradigms. EA + SI is a policy learning method that uses the semantic feature cognition (SFC) and the sequential cognitive memory (SCM) to process observations. EA is a policy learning method that uses the SFC to process observations. SI is a transformer-based method that uses the SFC and the SCM to generate semantic tokens. STF is an adaptive embodied learning method based on the student-teacher framework. NBV is a region-based viewpoint optimization computational method.
Back to article page