Prototype of wheelchair commanded by voice using HMM in a controlled environment

Abstract

This article presents the development of a recognition system of speaker independent isolated words to command a wheelchair. Each word is encoded using techniques of Linear Prediction and Real Cepstrum, while the classification step is performed using Hidden Markov Models (HMM). From the results, commands are generated to a mobile system (wheelchair), which is led in a controlled environment.

Keywords

• modelos Ocultos de markov
• procesamiento de voz
• predicción lineal
• Cepstrum