Pseudo-Random Binary Sequences Synchronizer Based on Neural Networks

Abstract

Modern telecommunication market grows rapidly. Regular user of the telecommunication networks gets higher data rates and higher quality of services every year. Inherent element of this rapid progress of the services is the need to develop faster and better devices for network testing. Very important role in the test environment play Pseudo-Random Binary Sequences (PRBS) generators and synchronizers. Their functionality is described by the logical equations from the years, so it seems to be useless to make new research about them. However, experiments described in this publication show something just opposite. Authors of this paper have improved design of the PRBS synchronizer with neural network and new protocol. Proposed implementation has ability to get faster in the synchronization state and it is more resistant to the transmission errors. Finally, neural network synchronizer have overall better parameters than classic solutions, what will be subject of presented research.