Comparison of different Neural Network architectures for SBTVD signal prediction based on effective transmission heights

In this article, we propose three different neural networks architectures for predicting the received signal power of the Brazilian digital TV system based on the effective transmission heights for each frequency channel and the points of reception inside the coverage area. The presented models will be analyzed and compared with the main propagation models used for digital TV described in the scientific literature. An accurate estimate of the received signal is essential to ensure a good image quality of the digital TV sets of all customers of the system.