Amauri Lopes

A solução clássica para o problema de estimação dos ângulos de chegada (DOA) de sinais incidindo em um arranjo de sensores é a aplicação do método de máxima verossimilhança. Este método leva ao problema de otimização de uma função custo não-linear, não-quadrática, multimodal e variante com a relação sinal-ruído (SNR). Os métodos propostos para tal tarefa, presentes na literatura, falham em uma ampla gama de valores de SNR. Este trabalho apresenta os resultados de um estudo sobre a aplicação de ferramentas pertencentes à computação natural ao problema de estimação DOA. Simulações demonstram que quatro dos algoritmos analisados alcançam o ótimo global para uma ampla faixa de valores de SNR, com esforços computacionais inferiores àquele exigido por uma busca exaustiva.

This paper describes for the first time, to our knowledge, the development of an Eye Diagram Analyzer equipment transparent to signal bit rate, based on the reconstruction of an eye diagram from a set of asynchronous undersampled data using digital signal processing techniques. It is intended to be used for optical network performance monitoring based on bit error rate estimation through Gaussian fitting Q-Factor calculation.

This paper presents the initial studies and the structure adopted in the development of a new method for objective assessment of audio quality, named Objective Measure of Audio Quality (Medida Objetiva da Qualidade de Audio-MOQA). New techniques are presented and their impact on the global performance of the method is analysed. The results are compared to that one reached by PEAQ method, which is currently adopted as standard by International Telecommunication Union.

We propose an estimator for the direction of arrival (DOA) of plane waves incident on a linear equally spaced array of sensors. The estimator uses a forward-backward linear prediction filter optimized by means of the total least squares criterion subject to constraints. The DOA angles are estimated by the zeros of the optimum filter. The proposed procedure is compared with two maximum likelihood methods presented in the literature.

This paper deals with the application of the Kohonen Self-Organizing Maps (KSOM) to methods of objective speech quality assessment. The performance of the objective methods so far proposed depends on many factors, among which the required mapping between the objective and subjective domains is one of the most important. The purpose of this paper is to present new results about application of the KSOM networks to replace the traditional third-order polynomial mapping. Some distinct network topologies and techniques for the extraction of the signal parameters are presented and tested in the context of the "Medida Objetiva de Qualidade de Voz" (MOQV) objective method applied to situations present in a traditional database.