Table 13 Comparison with other Models in Terms of Time Complexity, Loss and Accuracy.

From: CTDNN-Spoof: compact tiny deep learning architecture for detection and multi-label classification of GPS spoofing attacks in small UAVs

Channel

Model

Time (s)

Loss

Accuracy

Ch0_output

CTDNN-Spoof

12.3

0.0217

0.9485

Model 2

11.5

0.0170

0.9659

Model 3

10.8

0.0115

0.9612

Ch1_output

CTDNN-Spoof

13.1

0.0200

0.9503

Model 2

12.7

0.0250

0.9496

Model 3

11.4

0.0190

0.9356

Ch2_output

CTDNN-Spoof

14.0

0.0132

0.9806

Model 2

13.6

0.0033

0.9659

Model 3

13.2

0.0115

0.9966

Ch3_output

CTDNN-Spoof

15.5

0.0122

0.9739

Model 2

14.8

0.0123

0.9749

Model 3

14.1

0.0082

0.9712

Ch4_output

CTDNN-Spoof

16.7

0.0143

0.9766

Model 2

15.9

0.0078

0.9659

Model 3

15.2

0.0115

0.9842

Ch5_output

CTDNN-Spoof

17.3

0.0096

0.9858

Model 2

16.8

0.0022

0.9955

Model 3

16.0

0.0015

0.9946

Ch6_output

CTDNN-Spoof

18.5

0.0098

0.9828

Model 2

17.9

0.0048

0.9901

Model 3

17.2

0.0031

0.9898

Ch7_output

CTDNN-Spoof

19.0

0.0224

0.9314

Model 2

18.3

0.0269

0.9461

Model 3

17.6

0.0198

0.9335