DNN-based Beam and Blockage Prediction in 3GPP InH Scenario.

In this paper, we investigates DNN-based beam and blockage prediction in a millimeter-wave (mm-wave) indoor hotspot scenario. First,a deep neural network (DNN) is designed to learn the mapping between the user positions along with their data traffic demands and the corresponding blockage statuses and optimal beam indices. Following this, a fingerprinting database is created during an offline learning phase to train the proposed DNN, which consists of user positions along with their data traffic demands and their corresponding blockage statuses and optimal beam indices that maximize the reference signal received power via an exhaustive search. During a subsequent online learning phase, the trained DNN is utilized to predict the optimal tunings of beams and blockages corresponding to the targeted user locations with the given data traffic demands. System-level simulations are conducted to assess the accuracy of blockage prediction based on the 3GPP new radio channel and blockage models. The simulation results reveal that the proposed scheme is capable of predicting mm-wave blockages with an accuracy greater than 90%. Furthermore, these results confirm the viability of the proposed DNN model in predicting the optimal mm-wave beams and spectral efficiencies.