Time synchronization method based on ELM

Abstract

The invention discloses a time synchronization method based on ELM, and the method comprises the steps: collecting a received signal, and constructing a training data set i=1,2,...,Nt; constructing anetwork model based on ELM time synchronization; training the network model based on ELM time synchronization offline and storing the model and parameters; running the network model based on ELM timesynchronization on line to obtain a time synchronization estimation value. According to the time synchronization method disclosed by the invention, the time synchronization network model is constructed by virtue of the ELM network, and compared with a classical cross-correlation synchronization method, the time synchronization performance of the system can be effectively improved; when nonlinear distortion exists in the system, the time synchronization method is still applicable.

Description

technical field [0001] The invention relates to the technical field of time synchronization, in particular to a time synchronization method based on an ELM (extreme learning machine). Background technique [0002] There are a large number of nonlinear devices or modules in the transmitter of the communication system, which causes the inevitable nonlinear distortion of the transmitted signal. As one of the important components of wireless communication system, the synchronization performance of time synchronization is seriously degraded under nonlinear distortion conditions. [0003] The classical time synchronization scheme based on the "cross-correlation method" is due to the existence of nonlinear distortion, the orthogonality of the training sequence is destroyed, so the synchronization performance is greatly reduced, and then it is difficult to apply under the condition of nonlinear distortion; and the use of deep neural network learning Although time synchronization ca...

AI Technical Summary

Problems solved by technology

[0002] There are a large number of nonlinear devices or modules in the transmitter of the communication system, resulting in the inevitable nonlinear distortion of the transmitted signal

As one of the important components of the wireless communication system, the synchronization performance of time synchronization is seriously deteriorated under nonlinear distortion conditions

[0003] The classical time synchronization scheme based on the "cross-correlation method" is due to the existence of nonlinear distortion, the orthogonality of the training sequence is destroyed, so the synchronization performance is greatly reduced, and then it is difficult to apply under the condition of nonlinear distortion; and the use of deep neural network learning Although time synchronization can improve system time synchronization performance, it faces a series of problems such as complex parameter tuning and long training time

Images