A Multirate Sensor Information Fusion Strategy for Multitask Fault Diagnosis Based on Convolutional Neural Network

Table 2

Summary of the submodel of MRSIFS using 2D CNN for feature extraction.

Network structure	Layers	In_channels	Out_channels	Kernel_size	Stride	Padding
Input layer	Conv2D	7	16	7	2	3
	MaxPool			3	2	1
Downsampling layer	Conv2D	16	16	1	1	0
	Conv2D	16	16	3	1	1
	Conv2D	16	64	1	1	0
	Conv2D	64	16	1	1	0
	Conv2D	16	16	3	1	1
	Conv2D	16	64	1	1	0
Smooth layer	Conv2D	64	64	1	1	0
Upsampling layer	“Bilinear”
Output layer	Conv2D	64	32	1	1	0
	MaxPool			4	4	0
	Dropout ()