Rapidly Varying Channel State Estimation Employing Deep Learning Techniques for Predictive Analysis

With the gradual development of 5G technology and the in-depth research of 6G technology, many high-speed mobile application scenarios have put forward more stringent performance requirements for mobile communication networks. Specifically, the communication link must ensure extremely low error rates while meeting the dual demands of high communication quality and high transmission rate. In view of this, adaptive link adjustment technology has become the focus of industry research. However, one key foundation for achieving adaptive link technology is the ability to accurately sense channel state information (CSI). Given this, this paper adopts the deep learning technology of Transformer based on attention mechanism and combines the CNN-LSTM model based on neural network and LSTM network with fast-varying signal prediction to innovatively integrate it with Transformer technology for the first time, constructing a novel model for fast-varying channel prediction-CLTE (CNN-LSTM-based Transformer Prediction Model). Through experimental verification, the model achieves an average performance improvement of 0.8 to 1.5 in the mean square error (MSE) of SNR prediction value compared to the traditional CLSTM-based CSI prediction model. Consequently, the efficacy of adaptive link deployment within the context of rapid-channel prediction is set to be enhanced.