Abstract: In view of the problem that IoT terminals are vulnerable to interference in the identification process,a non-interference identification method of IoT terminals based on one-dimensional convolution neural network is proposed.Acceleration data for different gaits are collected through the built-in acceleration sensor of the Internet of Things terminal and preprocessed through smoothing and windowing processing.Gait data fusion model is used to analyze user data under different behavior-position combinations,and random forest model is used to predict the position of IoT terminals and target data behavior.Behavioral target data are integrated with positional data to obtain gait fusion data.On this basis,a one-dimensional convolutional neural network model is constructed,which takes the fused data as input and outputs the user's identity recognition results after iterative training.The experimental results show that the non-interference identity recognition method for IoT terminals based on one-dimensional convolutional neural networks can accurately identify the identity of IoT terminal users in appropriate model parameter conditions.

Key words: convolution neural network, IoT terminal, non-interference, identification, gait data, behavior-position combination