Hypertuning the parameters of spatial multiplexing using deep learning for improving the performance of OCN

Spatial multiplexing (SM) is considered in optical communication to improve the network’s performance. The SM divides the data signals into separate ones to transfer between the antennas. It helps optical communication networks better utilize available resources such as transmitters and receivers. It also reduces the antennas required to transmit data through the optical networks. Artificial intelligence has led to the accelerated growth of all fields and has significantly contributed to the development of communication technologies. Various communication networks like cellular, WAN, LAN, and long-distance communication networks have adopted AI-based optimization for the better utilization of the available resources and improved longevity of the devices. Thus, this paper has motivated me to develop an AI algorithm for tuning and optimizing the OCN parameters to improve the overall efficiency concerning long distances. This paper proposes a deep learning model to optimize spatial multiplexing in MIMO systems to improve optical communication networks. It uses an autoencoder model to optimize the data transmission at the transmitter and receiver ends and enhance the model’s overall performance. The refractive index error faced in the optical fibers is minimized through IFFT (inverse fast Fourier transform), and the signal quality is improved with low PSD. Thus, the proposed hyper-tuning of spatial multiplexing parameters through a deep learning model (HSMDL) provides better optimization for the optical networks. Optical parameters like energy requirements, communication distance, and data quality are all monitored. The transmitted data is optimized through several layers and encoded and mapped through spatial multiplexing. The proposed model is evaluated with simulation and compared with various existing models in terms of SNR, BER, time, quality of transmission, and network sustainability.