Adaptive Channel Coding for MC-CDMA Systems: A Deep Learning Approach with Attention-Enhanced LSTMs for Efficient Error Correction

The increasing demand for high-speed and reliable wireless communication has highlighted the need for more efficient channel coding techniques in MC-CDMA systems. Traditional methods such as UTTCM and Polar Codes face challenges in computational complexity and adaptability, limiting their performance in dynamic wireless environments.This paper introduces a deep learning-based channel coding approach using an Attention-Enhanced LSTM (ALSTM) model, designed to improve error correction efficiency while reducing processing overhead. Simulation results across AWGN, Rayleigh, and Rician fading channels show that the proposed model achieves lower BER and reduces computation time by 44% compared to UTTCM and 38% compared to Polar Codes. Additionally, the deep learning model maintains stable performance as the number of users increases, demonstrating better scalability for dense wireless networks. These findings suggest that AI-driven channel coding can significantly enhance MC-CDMA systems, offering a promising path for next-generation wireless networks and 6G applications.