Intelligent Propagation Prediction for FM Broadcasting Driven by Data and Models

Due to its mature technology, low cost, and broad coverage, frequency modulation (FM) radio broadcasting still holds irreplaceable value. To promote the technological advancement of FM broadcasting and enhance the accuracy of signal propagation predictions, we integrate a series of algorithms, such as the coefficient of variation, neural networks, and genetic algorithms, along with observational data, to establish an intelligent high-precision FM radio signal propagation prediction model driven by data and models for Beijing area. Experimental validation demonstrates that the proposed model significantly outperforms the compared models in terms of accuracy and effectiveness. The main contributions are as follows. First, a novel method for improving the predictive performance of legacy propagation models is proposed. Specifically, we introduce a data-and-model-driven modeling framework for FM broadcasting, which combines the coefficient of variation model with neural networks. The neural network is further optimized using a multivariant and multicrossover adaptive genetic algorithm. The proposed method can significantly improve the prediction accuracy of legacy models. Second, a comparison of prediction accuracy with advanced existing models such as ITU-R P.1546, P.2001, and ITM shows that the proposed model improves prediction accuracy by 68.31%, 74.75%, and 71.89%, respectively, thereby fully demonstrating the superiority of the intelligent model driven by data and models.