An Improved Pelican Optimization Algorithm Based Residual Recurrent Neural Network for Channel Estimation with Hybrid Precoder in MIMO-OFDM System

Massive multiple input multiple output (MIMO) technology stands as a cornerstone in the evolution of 5G networks. Using an abundance of degrees of freedom, MIMO systems exploit complex spatial characteristics to improve performance. Despite its potential, designing effective 5G MIMO communication systems faces major challenges, primarily in system modeling and channel estimation. Dynamic and complex propagation environments present significant obstacles in accurately representing system behaviour and estimating channel parameters. In order to estimate the channel in the wireless network an improved deep learning based Residual recurrent neural network (ResRNN) is implemented in this paper. Firstly, the least square estimator is used to estimate or calculate the historic channel responses of a pilot block. With these collective responses of the channel, we train the responses using the new deep learning based ResRNN, in which the weight of channel parameters are selected optimally using Improved pelican optimization algorithm (IPOA). Using the IPOA-based ResRNN hybrid precoder the present channel responses can be evaluated or estimated. By varying the length of pilot sequence and by varying the size of antennas at the transmitter and receiver, the metrics such as the bit error rate (BER), mean square error and throughput for (signal to noise ratio) SNR are evaluated. Simulation results show the developed method outperforms the conventional methods by attaining a BER of 10–5 with SNR is 30 dB wireless channel access. The performance of the proposed IPOA-DERNN method is compared with the conventional channel estimation methods.