| Popis: |
Novel 5G communication architectures can utilize mmWave systems for high bandwidth applications. To establish a reliable mmWave communication link, its vulnerability to blocking, signal degradation, and propagation loss needs to be overcome. Large antenna arrays mitigate these deficiencies by acquiring high directional gain through beamforming. Conventional beamforming techniques must have prior knowledge of the user’s current location for directional signal transmission. This incurs practical difficulty when a non-stationary user is communicating with the base station. To avoid this problem, a novel way of detecting user location, by incorporating vision data, leveraging deep learning is proposed. The detected location can then be given to a beam forming antenna to steer the beams according to the user location. |
