Abstract:
The Beyond fifth Generation (B5G) communication systems imposed several challenges on radio designers. For example, a machine is required to set up a call at a low Signal-to-Noise Ratio (SNR), as low as −10 dB, in the extended coverage mode. Moreover, only one receive antenna will be available, and virtually no frequency diversity. Such requirements present major challenges to maintaining timing and frequency synchronization. Carrier Frequency Offset (CFO) estimation is at the heart of these challenges. Different approaches have been proposed for CFO estimation such as maximum likelihood based on a cyclic prefix. Nevertheless, these methods remain limited in various ways. At the same time, Machine Learning (ML) techniques showed outstanding performance in several wireless communication problems. In this work, we propose an ML-based approach for CFO estimation in OFDM systems. Specifically, we propose a Gradient-Boosting Machine (GBM)-based solution to predict the CFO given the received Primary Synchronization Signal (PSS) and Secondary Synchronization Signal (SSS). Furthermore, we make our dataset available for public access to encourage other researchers to pursue this promising direction. We compare our results with different baseline models (i.e., artificial neural networks and support vector machines). The experimental results show that our model outperforms other baseline models due to its ensemble nature which enables ensemble models to obtain a better generalization behavior.