ML-Based Intelligent O-RAN Control in 6G Integrated Terrestrial and Non-Terrestrial Networks

The Open Radio Access Network (O-RAN) specification aims to define open, interoperable RAN interfaces with virtualized, intelligent RAN functions to support next-generation integrated Terrestrial and Non-Terrestrial Networks (T-NTN). The O-RAN specification introduces several options in which ML functional blocks can be distributed across several layers of the T-NTN infrastructure to enable intelligent and optimized RAN control mechanisms on different time scales. Machine Learning solutions can be adapted to T-NTN to enable efficient centralized and distributed intelligence solutions with specific performance requirements. We further include the possibility of considering multi-time-scale intelligence solutions over the NTN RAN through the distributed control architectures supported by the O-RAN technology.