No abstract

The availability of global and scalable tools to assess disabled pedestrian level of service (DPLoS) is a real need, yet still a challenge in today's world. This is due to the lack of tools that can ease the measurement of a level of service adapted to disabled people, and also to the limitation concerns about the availability of information regarding the existing level of service, especially in real time. This paper describes preliminary results to progress on those needs. It also includes a design for a navigation tool that can help a disabled person move around a city by suggesting the most adapted routes according to the person's disabilities. The main topics are how to use advanced computer vision technologies, and how to benefit from the prevalence of handheld devices. Our approach intends to show how crowdsourcing techniques can improve data quality by gathering and combining up-to-date data with valuable field observations.

No abstract

ResumenEn este trabajo se presentan los resultados del diseño y desarrollo de una red neuronal de tipo feed-forward en un sistema embebido, para identificar patrones asociados con la pronunciación de fonemas vocálicos del idioma español. Para el entrenamiento de la red fueron utilizados los coeficientes cepstrales de las frecuencias de Mel (MFCC), extraídos a partir de señales de audio provenientes de la pronunciación de las vocales abiertas (/a/, /e/ y /o/). La captura de nuevas muestras se realizó a través de un computador y un sistema embebido para la evaluación, comprobación y comparación del desempeño de la red neuronal. Se obtuvo un porcentaje de identificación correcta por encima del 98%. Esto indica que el error en la separación de clases fue inferior al 2% en la red neuronal al momento de evaluar patrones asociados a alguna de las tres vocales. Basado en los resultados obtenidos se concluye que la implementación de algoritmos de inteligencia artificial para tareas de clasificación en sistemas embebidos es factible, y presenta resultados similares a los que tendría el mismo sistema operando con los recursos de un computador. Palabras clave: red neuronal; señales de audio; sistema embebido; coeficientes cepstrales; reconocimiento del habla Recognition of Vowel Patterns by Implementing an Artificial Neural Network in an Embedded System AbstractIn this work the results of the design and development of a neural network type feed-forward in an embedded system, to identify patterns associated with the pronunciation of vowels of Spanish language are presented. For the training of the network Mel frequencies cepstrals coefficients extracted from audio signals related to the pronunciation of the open vowels (/ a /, /e/ and /o/) were used. The capture of new samples was performed by using a computer and an embedded system to evaluate, test and compare the performance of the neural network. A percentage of correct identification above 98% was obtained indicating that the error in the class separation was less than 2% for patterns obtained from any of the three vowels. Based on the results it is concluded that implementation of artificial intelligence algorithms for classification tasks in embedded systems is feasible, and presents similar results to those that would have the same system working with the resources of a computer.