Potentials of MIMO and Neural Networks in Industrial Cognitive Networks

摘要

Spectrum reutilization is a potential aspect in cognitive radio (CR) and has been employed recently in the shared spectrum sub-6 GHz. Such application has been supported by technologies that can manage the interference at other networks while providing acceptable services for the CR network itself. The multiple input multiple output (MIMO) technology brings benefits to CR with regard to the interference management making it a promising technology for Industry 4.0. In this paper, we propose a MIMO system for industrial CR networks. The proposed MIMO improves not only the spectral efficiency performance, but also the latency. Our model addresses industrial augmented reality application which requires high data rate and low latency. Specifically, the proposed scheme employs efficient precoding and deep learning for solving non-convex power optimization problem. The numerical results demonstrate outstanding performance in sake of our proposal.