Arcabouço de Aprendizado Estatístico para Caracterização Assintótica de Canais sem Fio

Felipe M Laburú, Flavio P. Calmon, Juliano Assine, José Cândido Silveira Santos Filho

DOI: 10.14209/sbrt.2022.1570824788

Evento: XL Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2022)

Keywords:

Abstract

Artificial intelligence tools have attracted great attention from the wireless community, which keeps reformulating problems in this area through the statistical learning framework. This work proposes the following problem: from a set of channel samples, to estimate the statistical behavior of essential performance metrics (e.g., error rate and outage probability) within the system's operating range. An effective solution is presented using deep learning techniques and evaluated in light of the solution via classical inference, conceptually optimal but impracticable. The analysis is arranged into representative numerical results that reveal how each parameter affects estimation.

Download