Abstract
Multi-Input Multi-Output Orthogonal Frequency Division Multiplexing (MIMO-OFDM) is the main air interface for fifth-generation wireless communication. Channel estimation (CE) is a key factor for attaining efficient system performance of 5G wireless networks. The Deep Learning (DL) approaches employed in MIMO-OFDM systems can minimize the computational complexity of 5G networks while increasing the system performance and reliability. In this research work, an Optimized Deep Multi-Scale Three-Dimensional CNN-dependent Channel Estimation in a 5G MIMO-OFDM (DMTDCNN-CE-5G-MIMO-OFDM) System is proposed. It provides a comprehensive analysis of DMTDCNN methods, 5G channel methods, DMTDCNN-based channel estimation, and error performance for MIMO-OFDM receivers. Here, channel estimation plays a main role in the effectual system performance on the 5G wireless networks. Furthermore, DMTDCNN optimized with the Modified Social Group Optimization Algorithm (MSGOA) can develop the system performance with reliability by decreasing the computational complexity of 5G communication systems. It combines DMDTCN-MSGOA with convolutional and polar codes, utilizing hard and soft decoders, improving the quality of channel estimation by reducing PER and PAPR attenuation. In this case, 1/2-rate polar coding with block lengths of 128, 256, and 1024 bits is used, and the outcome is 1, 2, and 8 MIMO-OFDM signals that rely on 4-QAM. The proposed DMTDCNN-CE-5G-MIMO-OFDM method attains 20.89%, 33.45%, 25.67% high throughput and 15.67%, 22.57%, 38.98% low mean square compared with the existing models, such as DL-based CE in 5G MIMO-OFDM schemes (DL-CE-MIMO-OFDM), low complexity learning-dependent channel estimation OFDM by online training (LCL-CE-OFDM), dual CNN-based CE for MIMO-OFDM schemes (DCNN-CE-MIMO-OFDM) methods, respectively.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call