Aiming at the problem of poor prediction accuracy of Channel State Information (CSI) caused by fast time-varying channels in wireless communication systems, this paper proposes a gated recurrent network based on experience replay and Snake Optimizer for real-time prediction in real-world non-stationary channels. Firstly, a two-channel prediction model is constructed by gated recurrent unit, which adapts to the real and imaginary parts of CSI. Secondly, we use the Snake Optimizer to find the optimal learning rate and the number of hidden layer elements to build the model. Finally, we utilize the experience pool to store recent historical CSI data for fast learning and complete learning. The simulation results show that, compared with LSTM, BiLSTM, and BiGRU, the gated recurrent network based on experience replay and Snake Optimizer has better performance in the optimization ability and convergence speed. The prediction accuracy of the model is also significantly improved under the dynamic non-stationary environment.

Aiming at the problem of low estimation accuracy under a low signal-to-noise ratio due to the failure to consider the “beam squint” effect in millimeter-wave broadband systems, this paper proposes a model-driven channel estimation method for millimeter-wave massive MIMO broadband systems. This method considers the “beam squint” effect and applies the iterative shrinkage threshold algorithm to the deep iterative network. First, the millimeter-wave channel matrix is transformed into a transform domain with sparse features through training data learning to obtain a sparse matrix. Secondly, a contraction threshold network based on an attention mechanism is proposed in the phase of beam domain denoising. The network selects a set of optimal thresholds according to feature adaptation, which can be applied to different signal-to-noise ratios to achieve a better denoising effect. Finally, the residual network and the shrinkage threshold network are jointly optimized to accelerate the convergence speed of the network. The simulation results show that the convergence speed is increased by 10% and the channel estimation accuracy is increased by 17.28% on average under different signal-to-noise ratios.

Aiming at the characteristics of resource periodicity in massive MIMO systems and bandwidth allocation without comprehensive consideration of user service QoS and channel state information, resulting in poor user satisfaction and low bandwidth utilization, this paper proposes an adaptive bandwidth allocation method based on user services. This method comprehensively considers factors, such as user service QoS, channel state information, and resource periodicity, to adaptively allocate bandwidth for users using different services. Firstly, based on the service priority, the user priority is dynamically adjusted according to the current channel state information and the continuous periodicity of the allocation, and the user is scheduled.; Secondly, the dynamic priority is combined with the minimum guaranteed time slot to establish the objective function of adaptive bandwidth allocation. Finally, chaos theory, Levy flight, and reverse learning are integrated to improve the bald eagle optimization algorithm. The improved bald eagle algorithm is used to solve the problem, and the optimal solution to bandwidth allocation is obtained. The simulation shows that compared with the traditional bandwidth allocation method based on user service quality perception, the bandwidth allocation algorithm based on the minimum rate requirement, and the ant colony-based allocation algorithm, the bandwidth allocation method proposed in this paper improves the system utility value, bandwidth utilization rate, throughput, and user satisfaction by 23.70%, 4.22%, 6.55%, and 4.28%, respectively, and better meets the business needs of users.

In order to solve the problem wherein too many base station antennas are deployed in a massive multiple-input–multiple-output system, resulting in high overhead for downlink channel state information feedback, this paper proposes an uplink-assisted channel feedback method based on deep learning. The method applies the reciprocity of the uplink and downlink, uses uplink channel state information in the base station to help users give feedback on unknown downlink information, and compresses and restores the channel state information. First, an encoder–decoder structure is established. The encoder reduces the network depth and uses multi-resolution convolution to increase the accuracy of channel state information extraction while reducing the number of computations relating to user equipment. Afterward, the channel state information is compressed to reduce feedback overhead in the channel. At the decoder, with the help of the reciprocity of the uplink and downlink, the feature extraction of the uplink’s magnitudes is carried out, and the downlink channel state information is integrated into a channel state information feature matrix, which is restored to its original size. The simulation results show that compared with CSINet, CRNet, CLNet, and DCRNet, indoor reconstruction precision was improved by an average of 16.4%, and outside reconstruction accuracy was improved by an average of 21.2% under all compressions.

A convolutional self-attention network-based channel state information reconstruction method is presented to address the issue of low reconstruction accuracy of channel state information in Multiple-Input Multiple-Output (MIMO) at a high compression rate. First, an encoder-decoder structure-based channel state information reconstruction model is built. The feature is extracted by the encoder’s convolutional network, and the information is compressed by adding an attention block. At the same time, the compressed information is nonuniform quantized to prevent the transmission process from using up too much bandwidth. A dequantization module and an attention block are added to the decoder to reduce the impact of noise on the matrix, converting the continuous value into a discrete value to increase reconstruction accuracy and using the long-time cosine annealing training approach. According to the simulation results, when compared to CsiNet, Lightweight CNN, CRNet, and CLNet, convergence speed is improved by 17.64%, indoor reconstruction precision is improved by an average of 37.4%, and outside reconstruction accuracy is improved by an average of 32.5% under all compressions.