Juliana C. Inácio, Vinícius J. D. Vieira, Guilherme C. Pereira, Renato Candido. Arbitragem de dispositivos acionados por voz em redes com nós heterogêneos. Simpósio Brasileiro de Telecomunicações – SBrT’2023, 2023, São José dos Campos, Anais do SBrT 2023, 2023, pp. 1–5. (in portuguese)

Abstract

Device arbitration occurs when there are devices equipped with smart virtual assistents activated by voice and by the same keyword in a same environment. Recently, deep-learning-based solutions were proposed with centralized processing and considering devices with identical characteristics. In this work, we propose an arbitration solution that uses classification models based on ensembles of decision trees, with descentralized processing and considering devices with distintic features. The proposed solution obtained an accuracy improvement up to 14% compared to the solution using energy-based arbitration.

Keywords

Device arbitration, source-localization, speech processing, machine learning.

Downloads

• PDF