Efficient PSS Detection Algorithm Aided by CNN

Abtract: In the 5G mobile communication system, cell search is the initial step in establishing downlink synchronization between user equipment (UE) and the base station (BS). Primary synchronization signal (PSS) detection is a crucial part of this process, and enhancing PSS detection speed can reduce communication latency and improve overall quality. This paper proposes a fast PSS detection algorithm based on the correlation characteristics of PSS time-domain superposition signals. Conducting PSS signal correlation within a smaller range can reduce computational complexity and accelerates communication speed. Additionally, frequency offset can impact the accuracy of calculations during the PSS detection process. To address this issue, we propose applying convolutional neural networks (CNN) for frequency offset estimation of synchronization signals. By compensating for the frequency of related signals, the accuracy of PSS detection is improved. Finally, the analysis and simulation results demonstrate the effectiveness of the proposed approach.

Keywords: 5G; CNN; cell search; PSS detection