A Power Efficient Scheme for Speech Controlled IoT Applications

Abstract

Speech recognition has been a subject of research since decades. Although it has wide applications in Artificial Intelligence and modern user interfaces, when speech processing is applied to embedded systems we also need to consider the constraints which are normally faced and algorithms to overcome the same. While the applications of embedded systems are now being massively focused on Internet of Things, still the prime research concerns are power efficiency and security. Here a state of the art scheme is proposed where speech processing is applied to the constrained IoT applications and the wireless communication is made power efficient. For achieving so, the 6LoWPAN protocol is implemented and Modified Vector Algorithm for Speaker Identification (MVA-SI) is used to increase PDF of the correct interpretation input speech through predefined wake up call.