Loading...
Publication
Relatórios automatizados de ocorrências públicas por meio de reconhecimento de voz
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 8.92 MB | Adobe PDF |
Advisor(s)
Abstract(s)
A presente dissertação explora o potencial da tecnologia Voice over Internet Protocol
(VoIP) integrada com mecanismos de automação de processos, através do desenvolvimento de
uma aplicação mock que simula cenários de reporte e gestão de incidentes. O objetivo não é
modernizar sistemas de comunicação em uso real, mas demonstrar, em ambiente controlado,
como a convergência entre VoIP, reconhecimento automático de fala (ASR), síntese de fala (TTS)
e modelos de linguagem pode criar fluxos de comunicação mais rápidos, precisos e menos
dependentes da intervenção humana.
A arquitetura proposta assenta no Asterisk, utilizado como PBX principal e integrado
com o UniMRCP para ligação a serviços de ASR e TTS, enquanto o GPT-4 assegura
processamento avançado de texto. A aplicação mock permite simular chamadas VoIP e
interações web, assegurando triagem automática de incidentes e geração de registos
estruturados.
Os resultados demonstram que a integração destas tecnologias reduz atrasos, elimina
inconsistências manuais e fornece dados organizados para análise em tempo real. Embora não
se destine a substituir sistemas críticos, o protótipo confirma o valor do VoIP como motor de
automação e inovação em comunicações digitais.
This dissertation explores the potential of Voice over Internet Protocol (VoIP) technology integrated with process automation mechanisms, through the development of a mock application that simulates incident reporting and management scenarios. The aim is not to modernize real-world communication systems but to demonstrate, in a controlled environment, how the convergence of VoIP, Automatic Speech Recognition (ASR), Text-to- Speech (TTS), and large language models can create faster and more accurate communication flows, with less human intervention. The proposed architecture is based on Asterisk, used as the main PBX, integrated with UniMRCP for ASR and TTS services, while GPT-4 provides advanced text processing. The mock application enables the simulation of VoIP calls and web interactions, supporting automatic incident triage and the generation of structured records. Results show that the integration of these technologies reduces delays, eliminates manual inconsistencies, and provides structured data for real-time analysis. Although not intended to replace critical systems, the prototype confirms VoIP’s role as a driver of automation and innovation in digital communications.
This dissertation explores the potential of Voice over Internet Protocol (VoIP) technology integrated with process automation mechanisms, through the development of a mock application that simulates incident reporting and management scenarios. The aim is not to modernize real-world communication systems but to demonstrate, in a controlled environment, how the convergence of VoIP, Automatic Speech Recognition (ASR), Text-to- Speech (TTS), and large language models can create faster and more accurate communication flows, with less human intervention. The proposed architecture is based on Asterisk, used as the main PBX, integrated with UniMRCP for ASR and TTS services, while GPT-4 provides advanced text processing. The mock application enables the simulation of VoIP calls and web interactions, supporting automatic incident triage and the generation of structured records. Results show that the integration of these technologies reduces delays, eliminates manual inconsistencies, and provides structured data for real-time analysis. Although not intended to replace critical systems, the prototype confirms VoIP’s role as a driver of automation and innovation in digital communications.
Description
Keywords
VoIP Process Automation Asterisk UniMRCP ASR GPT Automação de Processos