DEEP LEARNING-BASED BEAMFORMING COMMUNICATION SYSTEM AND METHOD

摘要

Provided are a deep learning-based beamforming communication system and method, wherein in an indoor environment using millimeter wave communication, in response to reference signals transmitted from a base station to at least one user terminal, reference signal received power and location information for each user terminal location are received from each user terminal and a fingerprint DB is constructed, and from the constructed fingerprint data, a user model is constructed on the basis of reference signal received power for each user terminal location and a blockage model is constructed on the basis of reference signal received power according to each blockage located between the base station and the user terminal. Location information and data traffic are received from the at least one user terminal, a beam index of the user terminal corresponding to the received data traffic is derived from a deep neural network, and a communication channel between the base station and a user is formed with the derived beam index, whereby reliability and a data transfer rate are improved in an indoor communication environment.