An SVR-Based Radio Propagation Prediction Model for Terrestrial FM Broadcasting Services in Beijing and Its Surrounding Area

For improving the accuracy and robustness of predicting radio propagation in the frequency bands of FM Broadcasting Services, we proposed a propagation prediction model suitable for multi-band and multi-scenario based on Support Vector Regression (SVR). The modeling is based on Beijing’s measurement data, covering various environments, such as rural, suburban, and urban areas, with a frequency range of 86.7MHz-106.6MHz. A comparison will be made to verify the model’s performance with the ITU-R P.1546 model, an internationally common model recommended by ITU. The results show that the model exhibits superior prediction accuracy and stability compared to ITU-R P.1546, and the average relative error of ITU-R P.1546 model prediction in each frequency band is prominent, from 3.01% to 4.77%. Moreover, the average relative error of each frequency band of the model is distributed by about 2%-2.5%. In addition, SVR’s the predicted direction and details are highly similar to ITU-R P.1546, which can better fit the trend of propagation loss with propagation distance and other conditions. As a result, the research can provide a reference for regional application and localization research of radio propagation prediction methods.