The article presents a modulation classification system for OFDM signals in Wi-Fi 6 and 5G networks. The key highlights are:
OFDM parameter estimation: The system estimates the OFDM symbol duration and cyclic prefix length using the cyclic autocorrelation function.
Feature extraction: The system extracts features characterizing the modulation of OFDM signals by removing the effects of synchronization errors. The features are converted into a 2D histogram of phase and amplitude.
Modulation classification: The 2D histogram is used as input to a convolutional neural network (CNN)-based classifier. The classifier can identify high-order modulations like 256QAM and 1024QAM without requiring prior knowledge of protocol-specific information.
Evaluation: The system is evaluated using both synthetic AWGN channel data and real-world measured over-the-air (OTA) datasets. It achieves a minimum accuracy of 97% with OTA data when the SNR is above the value required for data transmission.
Robustness: The system is designed to be robust to diverse OFDM signal configurations, including varying FFT size, cyclic prefix length, and carrier frequency, without requiring access to protocol-specific information.
In eine andere Sprache
aus dem Quellinhalt
arxiv.org
Tiefere Fragen