Prototype development of a smart voice-controlled audio system based on the Raspberry Pi platform

Abstract

We report a prototype audio system that can be controlled by means of voice commands. The operations of this audio player – such as song playback and volume control – can be effectively controlled by verbal commands alone without requiring any button-press or dial-turn, enabling a user to operate it with full hands. This feature could be particularly useful in situations when the user’s view and attention are pre-occupied, as in the case of driving or cycling. A voice-command device also has the advantage of a smaller form factor and simpler product casing, due to fewer dials and buttons. While there are a wide variety of commercial audio systems, only very limited few allow voice-control operations. This voice-controlled audio system is low-cost, open-source, and capable of “smart” features that include notifying weather forecasts, automatic volume control, and automatic shutdown. Its platform is the single-board computer Raspberry Pi, which runs speech-to-text processes that record and transcribe a given voice command, such as “Play music” and “Volume up.” The Raspberry Pi also connects to cloud-based services – and upon voice requests – can check and tell real-time information such as date, time, and weather. The system’s average response time to a given voice command was measured to be about 5 seconds. A simple circuit of LED indicators was also designed to prompt users to give voice commands at the right window of time, therefore improving system responsiveness and overall user experience. This prototype can be extended to voice-control other appliances in an integrated smart home environment.