Audio signal recognition in IoT using recurrent neural networks

The Internet of Things (IoT) is a concept that refers to the digital interconnection between ordinary objects through the Internet. Their growing presence in a wide range of applications and their growing computing and processing capabilities make them a valuable research area. IoT capabilities will revolutionize the interactions between physical objects and humans and the capacity collect useful data. Simultaneously advances in Artificial Intelligence have vastly changed the way computing devices process content of sequential data, such as images, video, voice, and audio. Recurrent Neural Networks are a powerful computational model that combines sequential data and Artificial Intelligence. The application of neural networks to IoT devices opens a new generation in Artificial Intelligence capable of performing complex detection and recognition tasks and human interactions with devices and their physical environment. This project studies the potential of using and developing a recurrent neural network architecture aimed at solving the problem of audio recognition in IoT, such as distinctive sounds that are produced by certain objects or actions in urban environments. Specifically, the approach of this project uses Recurrent Neural Networks programmed in Python language together with a USB microphone and the Raspberry Pi computer to form a low-cost IoT system able to recognize urban sounds.

• Κύριος συγγραφέας: Cortez Castro, Ronny Leonardo (author)
• Μορφή: bachelorThesis
• Γλώσσα: eng
• Έκδοση: 2022
• Θέματα: Redes neuronales recurrentes; Internet de las cosas; Raspberry Pi; Reconocimiento de audio; Inteligencia artificial; Recurrent neural networks; Internet of things
• Διαθέσιμο Online: http://repositorio.yachaytech.edu.ec/handle/123456789/537