CNN-LSTM-assisted joint channel estimation and signal detection for optical wireless communications

Area of Science:

• Optical Wireless Communication
• Signal Processing
• Machine Learning

Background:

• Multi-input multi-output (MIMO) orthogonal frequency division multiplexing (OFDM) optical wireless communication (OWC) systems face performance degradation due to channel fading and atmospheric turbulence.
• Accurate channel state information (CSI) is vital for reliable channel estimation but often necessitates high pilot overhead.

Purpose of the Study:

• To propose a joint channel estimation and signal detection (JCESD) algorithm to reduce pilot overhead and improve detection performance in MIMO-OFDM OWC systems.
• To leverage deep learning, specifically convolutional neural network (CNN) and long short-term memory (LSTM) networks, to learn channel characteristics and mitigate signal processing complexity.

Main Methods:

• Developed a joint channel estimation and signal detection (JCESD) algorithm integrating CNN and LSTM networks.
• Applied layer normalization during data pre-processing to improve performance.
• Evaluated the algorithm's performance using simulations in varying atmospheric turbulence conditions.

Main Results:

• The JCESD algorithm demonstrated superior performance compared to traditional LS-ZF and LS-MMSE algorithms, achieving up to a 1 dB gain at a 10^-5 bit error rate (BER) with reduced pilot overhead.
• The algorithm maintained performance with fewer pilots (e.g., 0.9 dB gain with four pilots versus eight).
• Achieved superior robustness in atmospheric turbulence channels with manageable computational complexity.

Conclusions:

• The proposed JCESD algorithm effectively reduces pilot overhead while enhancing signal detection performance in MIMO-OFDM OWC systems.
• The integration of CNN and LSTM networks provides a robust solution for channel estimation and signal detection under atmospheric turbulence.
• The JCESD algorithm offers a practical approach for improving the reliability and efficiency of optical wireless communication systems.