Service discovery and session establishment for machine-learning-based beam prediction in wireless communications

Abstract

Aspects of dynamic and interactive machine learning and feature extraction techniques for performing beam interference management are disclosed. In one aspect, upon entering a coverage area, a UE may transmit a discovery message including one or more machine learning (ML) services for ML service discovery. Based on the ML service discovery, the UE may transmit a session request to establish a data service session between the UE and a network node based on the ML service discovery and other criteria such as extracted features. The network node may receive the ML discovery data and extracted features and aggregate this information with other intelligent network devices to enable the network node to predict a beam blockage during the ML inference data service session. The network node can adapt beam blockage predictions by changing the timing and direction of communications between network entities.