INTEGRATION OF SPEECH RECOGNITION IN IOT-BASED SMART HOME AUTOMATION SYSTEMS

Abstract

This research presents the design and development of an Internet of Things-based Smart Home system that allows users to control household electronic devices using voice commands. The system utilizes an ESP32-S3 microcontroller integrated with an INMP441 digital microphone and Deepgram speech recognition service through a cloud-based intermediary server. The microphone captures user speech and sends digital audio data to the ESP32-S3 via an I2S interface, which then sends it to the server for real-time conversion into text. The recognized text is sent back to the ESP32-S3 to execute specific commands that activate or deactivate electronic devices through a four-channel relay module. Experimental results show that the system can recognize Indonesian voice commands with an average accuracy of 80 to 90 percent, with a response time of approximately one second and effective performance within a distance of three meters at a noise level of 40 to 55 decibels. This system provides a convenient, efficient, and modern way for users to control household appliances without physical contact or additional applications. It also offers significant benefits for the elderly or people with disabilities. Overall, the proposed design demonstrates a practical implementation of Internet of Things-based voice-controlled Smart Home technology, which improves convenience, accessibility, and energy efficiency in everyday home automation.