الفهرس | Only 14 pages are availabe for public view |
Abstract The increasing demands for a high data rate wireless connectivity necessitate the search for suitable spectrum regions to meet anticipated future requirements. The research community has shown considerable interest in millimeter-wave (mmwave) communication. This thesis presents a survey of different precoding or beamforming techniques that have been proposed in the literature. Hybrid beamforming techniques that combine analog and digital precoding can be adopted for mm-wave massive mimo wireless systems to minimize the power consumption and hardware complexity. The performance of the most common hybrid precoding algorithms has been investigated in this thesis. A convolution neural network (CNN) structure based on deep learning technique are suggested in this thesis. It can be trained to optimize and maximize spectral efficiency and it is compared with the traditional techniques. Also, a comparison has been made between the proposed technique and many deep learning techniques currently in use for detecting hybrid precoding and combining. It has been found that, DarkNet-19 gives the best performance for detecting hybrid precoding and combining. |