LTE-eMBMS systems efficiently deliver multicast/broadcast services using Layered Division Multiplexing (LDM) technology. In a two-layer LDM system, Layer 1, with higher power allocation delivers mobile services, and Layer 2 in a Single Frequency Network scheme provides local content. The challenge is to reduce the gap in the layers’ coverage areas caused by the use of different constellations, and SFN gain for Layer 2. Hence, the precision in the coverage area estimation is crucial for the successful planning and deployment, particularly regarding the SFN gain contribution in Layer 2. For this purpose, a real digital TV broadcasting SFN system was used as a model to design a method based on Machine Learning algorithms, aiming to enhance the coverage area precision for the Layer 2 in eMBMS. The method is able to estimate SFN gain value with a Mean Absolute Error (MAE) of 0.72 dB and certainty in positive or negative contribution in 93% of the cases.

中文翻译：

---

神经网络的 SFN 增益预测用于增强 LDM 系统中的第 2 层覆盖

LTE-eMBMS 系统使用层分复用 (LDM) 技术有效地提供多播/广播服务。在两层 LDM 系统中，具有较高功率分配的第 1 层提供移动服务，而单频网络方案中的第 2 层提供本地内容。挑战在于减少由于使用不同星座而导致的各层覆盖区域的差距以及第 2 层的 SFN 增益。因此，覆盖区域估计的精度对于成功的规划和部署至关重要，特别是对于 SFN在Layer 2中获得贡献。为此，以真实的数字电视广播SFN系统为模型，设计了一种基于机器学习算法的方法，旨在提高eMBMS中Layer 2的覆盖区域精度。