A good wireless network design depends on technical, financial and what criteria you want to meet. New technologies and services, such as transmission in 5G, and the reuse of frequencies, several new studies are being done to ensure the best project for a given area. This work presents a discrete radio propagation model that uses the K-nearest neighbors classifier. The model takes into account different attributes of an environment, the studied environment was in the dependencies of the Federal University of Pará (Pará-Brazil), representing a typically Amazonian environment. For design the model, measurement campaigns were carried out in three different frequencies: 521MHz, 2100 MHz, and 2600 MHz. The study of the fading in these frequencies was made to make a generalization of frequencies to include in the model. Two scenarios to optimize the positioning of the base stations under study were carried out. The proposed model was applied in some scenarios, showing the power received at each point, in addition, tests for frequency of 700 MHz were performed to show how good the polynomial obtained was in the generalization of the frequencies.
|
