Full-Duplex Voice Control for Electronic Devices

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current voice control systems for electronic devices require frequent wake-up operations, leading to complex user interactions and disrupted functionality, particularly in non-chat modes like music or video playback, resulting in a poor user experience.

Innovation Solution

Implementing a full-duplex listening state after entering a non-chat mode allows for direct identification and execution of voice commands without wake-up words, ensuring continuous operation and improved user experience by maintaining the listening state for a predetermined time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the intelligent device requires wake-up operations for each voice interaction, then the device can accurately identify user commands, but the operation becomes complicated and normal work is disrupted

Engineering Contradiction:
Improvevoice command identification accuracyVSAvoiduser interaction complexity
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The system performs preliminary wake-up word detection before entering full voice interaction mode. The wake-up word detection module continuously monitors for wake-up words in advance, and upon detection, activates the voice interaction module to handle subsequent commands without requiring additional wake-up operations, thus resolving the contradiction between accurate identification and operational simplicity

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts its listening state between wake-up word detection mode and full voice interaction mode. The voice interaction module is activated only when needed (after wake-up word detection) and deactivated afterward, creating a dynamic system that balances accuracy with operational simplicity by being highly active during interaction and low-power during standby

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If the intelligent device is woken up for each voice command, then the device can process commands accurately, but the current working state stops and functionality is disrupted

Engineering Contradiction:
Improvecommand processing accuracyVSAvoiddevice working continuity
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The wake-up word detection module operates continuously in the background before full voice interaction is needed, preparing the system for upcoming commands. This preliminary detection allows the device to maintain its current working state while still being ready to accurately process voice commands when the wake-up word is spoken, thus resolving the contradiction between processing accuracy and working continuity

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system maintains continuous wake-up word detection capability without fully activating the voice interaction module. The wake-up word detection module runs continuously to monitor for activation signals, allowing the device to maintain other productive tasks while being ready to switch to voice interaction mode when needed, ensuring both accuracy and continuity of useful actions

Inventive Principle:
Principle #20Continuity of useful action

3Reliability

If wake-up operations are required before voice interaction, then the device can distinguish valid commands, but the operation cost increases and user experience deteriorates

Engineering Contradiction:
Improvecommand validation reliabilityVSAvoidtime to initiate voice command
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs wake-up word detection in advance as a preliminary step before full voice command processing. This preliminary detection is continuously monitored and ready to activate the voice interaction module immediately when triggered, thus maintaining reliable command validation while minimizing the time users need to wait to initiate commands

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically transitions from a low-power wake-up word detection state to a high-performance voice interaction state only when needed. This dynamic switching reduces the average time cost by keeping the system in a ready-but-low-power state most of the time, while ensuring reliable validation through the continuous wake-up word monitoring capability

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3745252B1Voice control method and apparatus of electronic device, computer device and storage medium
Publication Date: 2023.10.25 BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
  • EP3745252B1 patent drawingFigure 1
  • EP3745252B1 patent drawingFigure 2
  • EP3745252B1 patent drawingFigure 3

AI summary

The present disclosure provides a voice control method and apparatus of electronic device, a computer device and a storage medium, the method includes: receiving (101; 201) a first voice command of a user, and entering a non-chat mode according to the first voice command; initiating (102) a full-duplex listening state of the electronic device, in which under the full-duplex listening state the electronic device identifies a first-type user command, the first-type user command having no wake-up word; acquiring (103) the first-type user command under the full-duplex listening state, and identifying the first-type user command; when the first-type user command is a preset command, executing (104) the first-type user command.