Skip to main content

Table 2 Results from training data of fourth and fifth sessions

From: Brain–machine interface based on deep learning to control asynchronously a lower-limb robotic exoskeleton: a case-of-study

   

A01

A02

A03

A04

A05

Avg.

Static

DL

Re-train all

50.8

55.9

67.6

76.5

69.9

64.1 ± 10.6

Re-train last 3

43.5

55.2

64.8

74.2

68.5

61.3 ± 12.1

No re-train

51.7

53.8

62.9

70.5

65.1

60.8 ± 7.9

FA

55.4

60.1

57.3

65.9

64.5

60.6 ± 4.6

Motion

DL

Re-train all

52.9

66.2

61.7

65.2

54.2

60.1 ± 6.2

Re-train last 3

49.2

59.7

50.6

60.1

49.0

53.7 ± 5.7

No re-train

53.3

61.0

52.4

60.6

51.7

55.8 ± 4.6

FA

56.5

59.8

55.5

57.8

55.1

56.9 ± 1.9

  1. DL deep learning, FA features-based approach