High Frequency Injection for False Acceptance Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Information handling systems often incorrectly process audio signals intended for other devices, leading to unintended actions, such as smart home devices responding to TV advertisements, due to the inability to distinguish between user-generated voice commands and background audio.
Innovation Solution
Incorporating a frequency signal injection method where an information handling system adds an identification signal to audio data, allowing receiving systems to identify and attenuate unwanted audio signals, thereby preventing false triggering of voice-activated functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If microphones receive all audio signals including background audio, then the system can capture all voice inputs, but false triggering of voice commands occurs
Solution Approach 1:
An identification signal is introduced as an intermediary element between the audio source and the voice command processing system. This signal is injected into the audio output of devices like TVs and smartphones, and detected by the receiving system to determine whether the audio contains legitimate voice commands or background audio that should be ignored.
Solution Approach 2:
The system changes the parameter of the audio signal by injecting a high-frequency identification signal (e.g., 19 kHz or 20 kHz tone) that is inaudible to humans but detectable by microphones. This parameter change allows the receiving system to distinguish between different types of audio sources without affecting human perception.
2Productivity
If the system processes all audio data without filtering, then no voice commands are missed, but processing time and computational resources increase
Solution Approach 1:
The receiving system performs preliminary action by detecting the identification signal before full voice command processing occurs. This early detection allows the system to quickly determine whether further processing is necessary, avoiding wasted computational resources on background audio.
Solution Approach 2:
The system extracts and processes only the identification signal component from the audio data stream separately from the voice command content. This extraction allows for rapid filtering decisions without the need to analyze the entire audio signal in detail.
3Reliability
If an identification signal is injected into audio output, then false triggering is reduced, but the audio signal complexity increases
Solution Approach 1:
The identification signal is injected at specific local points in the audio processing chain (at the audio output of transmitting devices) and detected at specific local points (at the microphones of receiving devices). This localized approach minimizes the impact on overall system complexity while achieving the desired false acceptance reduction.
Data Source
AI summary
Methods and systems for high frequency injection and detection for improved false acceptance reduction are disclosed. An information handling system may be configured to receive audio data and to add an identification signal to the audio data, wherein the identification signal is determined based on the audio data. The combined audio data and the identification signal may be output to a receiving device. An information handling system may also be configured to receive data that includes audio data and an identification signal that is associated with one or more frequencies in the audio data, identify the one or more frequencies in the audio data that are associated with the identification signal, and attenuate the one or more frequencies in the audio data to obtain modified audio data. The modified audio data may be output for audio processing.