Environment-aware intelligent numerology control approach for 5G and beyond systems

5G and beyond (5GB) systems show a flexible characteristic to meet different user communications requirements in parallel. Multi-numerology waveform design is an important part of this human centric characteristic, however, it is necessary to make multi-numerology planning. In this paper, the integrated sensing and communications (ISAC) concept is used with machine learning (ML) techniques to develop an intelligent numerology control mechanism for 5GB. Thanks to the ISAC capabilities, an environment-aware system is designed, and several useful sensing information (USI) that affect the wireless channel characteristic is utilized. It is assumed that the USI is obtained through 5GB smart city networks. At that point, synthetic data generation is performed to form a new dataset with USI-based features. Additionally, a novel feature control algorithm is proposed in the paper. Lastly, simulation results are presented for different type of ML models including several ensemble methods.