Touch Speech Time Window for Wake Word Elimination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for interactive displays face inefficiencies due to the need for wake words and rigid touch events to initiate speech recognition, which increases user interaction overhead and limits flexibility in dynamic environments with complex multimodal commands.
Innovation Solution
A touch and speech input system that utilizes a touch speech time window comprising a pre-touch, touch, and post-touch window, eliminating the need for wake words by actively listening for voice commands within a dynamic context, allowing flexible voice and touch interactions without specific triggering events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If wake words are used to trigger speech recognition, then the system can initiate voice commands, but the user interaction overhead increases due to the additional word that must be spoken
Solution Approach 1:
The system performs preliminary action by continuously monitoring audio input and maintaining readiness to execute commands without requiring a wake word trigger. The speech recognition system is pre-configured to recognize specific command phrases and execute them directly, eliminating the need for preliminary wake word activation.
Solution Approach 2:
The invention extracts and removes the wake word requirement from the command execution workflow. By separating the command recognition function from the wake word trigger, the system allows users to speak only the necessary command phrases without additional overhead words, directly reducing interaction steps.
2Adaptability or versatility
If rigid touch events are required to initiate speech recognition, then the system can trigger voice commands, but the flexibility in dynamic environments is limited
Solution Approach 1:
The system implements dynamics by making the speech recognition trigger adaptable to different contexts and user actions. Instead of requiring fixed touch events at specific locations, the system can be activated by various input methods including touches anywhere on the display, voice activations, or other dynamic user interactions, allowing flexibility in changing environments.
Solution Approach 2:
The invention applies universality by designing a speech recognition system that can be triggered through multiple different input methods and contexts. The system accepts various activation signals including but not limited to touch events, making it versatile across different usage scenarios and device states without requiring specific rigid triggers.
3Productivity
If speech recognition is continuously active, then commands can be executed immediately, but energy consumption increases
Solution Approach 1:
The system implements periodic action by activating continuous speech recognition only when needed based on contextual cues such as detected user presence, device state, or recent interaction patterns. Between these periodic activation periods, the system enters a lower-power state while maintaining the ability to quickly reactivate, thus balancing immediate command execution with energy conservation.
Data Source
Figure 1a
Figure 1b
Figure 2
AI summary
Method, apparatus, and computer-readable media for touch and speech interface includes structure and/or function whereby at least one processor: (i) receives an input from a touch sensitive input device; (ii) establishes a touch speech time window with respect to the received touch input; (iv) receives an input from a speech input device; (v) determines whether the received speech input is present in a global dictionary; (vi) if the received speech input is present in the global dictionary, determines whether the received speech input has been received within the established touch speech time window; and (vii) if the received speech input has been received within the established touch speech time window, activate an action corresponding to both the received touch input and the received speech input.