Speech Controlled IoT System Based on Context-Driven Rule Engine

Abstract

Smart systems are designed to improve various aspects of human activities, both in business and private environments. In most cases, such systems require a real-time insight into state and events in the given environment. We present a concept of a framework which can be used as a foundation for more advanced, experimental applications. The core framework provides generic, configurable monitoring and control mechanisms, and enables task automation based on events in the environment. Additionally, it includes a set of features allowing context sharing according to user-defined rules. Regarding possible applications, our emphasis was on speech interaction between human users and the system. Therefore, an experimental application is presented, providing recognition of simple spoken commands, with the ability to generate spoken notifications as well. The application is designed to support simple spoken interaction between a human user and the system in Croatian language, which makes a rare example.