Man-Machine Dialogue Mode Switching for TTS Misrecognition Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing man-machine dialogue systems face issues with misrecognition and unnatural interaction due to inappropriate dialogue mode selection, particularly in full-duplex and half-duplex modes, leading to disrupted conversations.

Innovation Solution

A method and device for automatically switching dialogue modes based on determining the dialogue field of user sentences, adapting to the most suitable mode for smooth interaction, including full-duplex or half-duplex transitions based on preset dialogue fields and user characteristics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If full-duplex dialogue mode is used, then natural interaction without audio missing is achieved, but misrecognition and false triggering occur due to TTS broadcast being recorded

Engineering Contradiction:
Improvenatural interactionVSAvoidrecognition accuracy
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system dynamically switches between full-duplex and half-duplex dialogue modes based on the dialogue state. When TTS broadcast is detected, the system transitions to half-duplex mode to prevent recording of broadcast audio, and switches back to full-duplex mode when broadcast completes, allowing natural interaction without misrecognition

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the recording parameter based on dialogue state. In full-duplex mode, recording is enabled for natural interaction; when TTS broadcast occurs, the system disables recording to prevent false triggering, thus changing the recording parameter adaptively to resolve the contradiction

Inventive Principle:
Principle #35Parameter changes

2Reliability

If half-duplex dialogue mode is used, then false triggering is avoided, but users must wait for broadcast completion causing unnatural interaction

Engineering Contradiction:
Improvedialogue stabilityVSAvoidinteraction naturalness
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system dynamically adjusts the dialogue mode based on real-time detection of TTS broadcast state. When broadcast is detected, it switches to half-duplex for stability; when broadcast ends, it switches to full-duplex for natural interaction, thus adaptively resolving the contradiction between stability and naturalness

Inventive Principle:
Principle #15Dynamics

3Productivity

If recording is kept on during TTS broadcast, then continuous dialogue flow is maintained, but AEC technology limitations cause misrecognition

Engineering Contradiction:
Improvedialogue continuityVSAvoidvoice recognition precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system changes the recording parameter from enabled to disabled when TTS broadcast is detected, and restores it after broadcast completes. This parameter change maintains dialogue continuity while avoiding misrecognition during broadcast periods when AEC technology is insufficient

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4054111B1Method for switching between man-machine dialogue modes
Publication Date: 2025.09.03 AISPEECH CO LTD
  • EP4054111B1 patent drawingFigure 1
  • EP4054111B1 patent drawingFigure 2
  • EP4054111B1 patent drawingFigure 3~4

AI summary

The present disclosure discloses a man-machine dialogue mode switching method, which is applicable to an electronic device. The method includes receiving a current user sentence spoken by a current user; determining whether a dialogue field to which the current user sentence belongs is a preset dialogue field; if yes, switching the current dialogue mode to a full-duplex dialogue mode; and if not, switching the current dialogue mode to a half-duplex dialogue mode. In the present disclosure, the dialogue mode is switched by determining whether the dialogue field to which the current user sentence belongs is the preset dialogue field, and the dialogue mode can be automatically switched and adjusted according to the difference of the dialogue fields, such that the man-machine dialogue is always in the most suitable dialogue mode and can be realized smoothly.