In order to solve the problems of low precision and low efficiency in load recognition of fiber-optic tactile sensors applied to flexible skin

a load recognition method for fiber Bragg grating(FBG)sensing array based on convolutional neural network(CNN)is proposed.Firstly

the sensing array is manufactured using Flexible resin and 3D printing technology.Secondly

points are divided and data are collected.Finally

CNN is used for recognition and compared with back propagation neural network(BPNN)and random forest(RF).The results show that the sensing array has good sensitivity and linearity

and CNN has good adaptability and generalization ability.In the recogniti

on of load position and magnitude

the evaluation indices root mean square error(RMSE)and coefficient of determination(

R

2

)are 0.4869 and 0.99477

respectively

which are better than those of the other two algorithms.This method can be applied to load sensing of flexible tactile sensors.