User State Prediction Based on Channel Quality Feedback in 5G Broadcast

With the rapid development of mobile wireless communication technology, the proliferation of devices in 5G networks necessitates enhanced resource management and user equipment coordination. As a promising solution, the operators can optimize the network management and resource allocation based on user status prediction in order to elevate the system efficiency and enhance the user experience. Under the 5G New Radio paradigm, one certain user equipment can feedback its current channel quality information via the uplink either periodically or non-periodically, and base stations could thereby identify static or dynamic users and optimize network management to adapt to user behavior. This feedback mechanism facilitates effective sensing of multicast channel quality and supports the decision-making processes involved in Modulation and Coding Scheme (MCS) selection and transmission mechanism determination. In this work, we leverage the Light GBM approach to forecast user state based on several uplink information, namely, Channel Quality Indicator (CQI), Reference Signal Receiving Power (RSRP), Reference Signal Receiving Quality (RSRQ) and Signal-to-Noise Ratio (SNR), to aid base stations in making well-informed transmission method selections in the future 5G broadcasting scenario. The result shows that our proposed method outperforms traditional Support Vector Machine (SVM) and Logistic Regression models in terms of user state classification accuracy.