A Hybrid Network Integrating MHSA and 1D CNN–Bi-LSTM for Interference Mitigation in Faster-than-Nyquist MIMO Optical Wireless Communications

• DOI: 10.3390/photonics11100982
• Publication Date: 2024-10-22T08:10:14Z
• Authors: Minghua Cao; Qing Yang; Genxue Zhou; Yue Zhang; Xia Zhang; Huiqin Wang
• Keywords: Nyquist–Shannon sampling theorem

To mitigate inter-symbol interference (ISI) caused by Faster-than-Nyquist (FTN) technology in a multiple input output (MIMO) optical wireless communication (OWC) system, we propose an ISI cancellation algorithm that combines multi-head self-attention (MHSA), one-dimensional convolutional neural network (1D CNN), and bi-directional long short-term memory (Bi-LSTM). This hybrid extracts data features using 1D CNN captures sequential information with Bi-LSTM, while incorporating MHSA to comprehensively reduce ISI. We analyze the impact of antenna numbers, acceleration factors, wavelength, turbulence intensity on system’s bit error rate (BER) performance. Additionally, compare waveform graphs amplitude–frequency characteristics FTN signals before after processing, specifically comparing sampled values four-pulse-amplitude modulation (4PAM) those obtained cancellation. The simulation results demonstrate within Mazo limit for selecting our proposal achieves 7 dB improvement BER compared conventional systems without deep learning (DL)-based algorithms. Furthermore, employing point-by-point elimination adaptive pre-equalization algorithm, exhibits comparable performance orthogonal transmission reducing computational complexity 31.15%.