This paper presents a hybrid deep learning-based channel equalization framework for MIMO-OFDM wireless communication systems. A Convolutional Neural Network combined with a Long Short-Term Memory (CNN-LSTM) architecture is proposed to effectively compensate for nonlinear channel distortions and inter-symbol interference under varying and complex channel conditions. The model is trained using simulated MIMO-OFDM data generated under Rayleigh fading environments, ensuring robustness across diverse signal scenarios. Performance evaluation is carried out using key metrics such as symbol error rate (SER), mean square error (MSE), and eye diagram analysis to assess signal clarity and distortion levels. Simulation results demonstrate that the proposed CNN-LSTM equalizer achieves significantly lower SER and MSE compared to conventional equalization methods, particularly at low and moderate signal-to-noise ratio (SNR) levels. These findings confirm that deep learning-based equalization offers improved robustness, adaptability, and enhanced performance for modern high-data-rate wireless communication systems.
Sachithanandam et al. (Tue,) studied this question.