Recognition of complex surfaces based on multiscale temporal networks

A review of robot grasping tactile perception based on deep learning

“…Gao et al [29] proposed the Multiscale Temporary Network (MTCN) architecture for robot tactile perception and recognition. The core module of the MTCN is the Temporary Convolutional Network (TCN) [30], chosen over LSTM [31] due to its advantages in hardware deployment and memory expansion.…”

“…(1) (2) (3) Figure 10. (1) The architecture of the TCN module, (2) the scheme of the residual block, (3) the structure of the SE block [29].…”