US20240161740
2024-05-16
Physics
G10L15/22
A method for managing multiple digital assistants on a single device involves receiving unique active sets of warm words for each assistant. These warm words are specific commands that the respective digital assistant can execute. The process includes a warm word arbitration routine that selects a final set of warm words from these active sets, allowing for efficient detection and execution of commands based on user utterances.
In environments equipped with connected microphones, users can issue voice commands to digital assistants without needing any physical devices nearby. For example, a user cooking in the kitchen can simply say, “please set a timer for 20 minutes,” and the assistant will respond accordingly. The presence of multiple assistants can lead to overlapping commands, necessitating a system to determine which assistant should respond to a user's request.
The system adapts dynamically to changes such as the addition or removal of digital assistants, or modifications to their respective warm words. It can also respond to shifts in ambient context, optimizing the detection process based on available resources and performance metrics like detection frequency and timing.
Users can configure which digital assistants operate simultaneously through voice commands or software requests. The system ensures that both assistants are ready to respond to user inputs, facilitating seamless interaction within multi-assistant environments. This flexibility enhances user experience by allowing multiple digital assistants to work together effectively.