Command Keyword Engine for False Positive Reduction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional wake-word engines in network microphone devices are prone to false positives due to false wake word triggers, which can lead to resource consumption and interrupt audio playback, as they often mistakenly invoke voice assistant services when detecting phonetically similar words or background audio.

Innovation Solution

Implementing a command keyword engine that detects specific command keywords only when certain playback conditions are met, allowing for local natural language processing and reducing false positives by not requiring data transmission to a voice assistant service, and using a local natural language unit to recognize a limited library of keywords for voice commands.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a traditional wake-word engine is used to detect voice inputs, then the device can invoke voice assistant services, but false positives occur due to phonetically similar words or background audio

Engineering Contradiction:

Improvevoice assistant service invocationVSAvoidfalse positive detection

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent segments the keyword detection function into two separate engines: a wake-word engine for invoking voice assistant services and a command keyword engine for media playback control. This segmentation allows each engine to be optimized for its specific function, with the command keyword engine focusing solely on playback commands without the false positive issues affecting wake-word detection

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The command keyword engine acts as an intermediary between audio input and media playback control, filtering and interpreting commands locally before executing actions. This intermediary layer processes audio locally without transmitting to cloud services, reducing false positives while maintaining reliable playback control

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of operation

If wake-word detection is continuously monitored, then voice assistant services can be invoked, but resource consumption increases and audio playback is interrupted

Engineering Contradiction:

Improvevoice input responsivenessVSAvoiddevice resource consumption

Core Design Contradiction:

Ease of operationVSUse of energy by moving object

Solution Approach 1:

Instead of continuous monitoring, the command keyword engine uses periodic action by only activating keyword detection when playback conditions are met (e.g., when media is actively playing). This allows the system to remain responsive to voice commands while conserving resources during periods when playback control is not needed

Inventive Principle:

Principle #19Periodic action

Solution Approach 2:

The system dynamically adjusts its operational state based on playback conditions. The command keyword engine enables or disables detection capabilities according to whether media playback is active, allowing the device to be highly responsive when needed while minimizing resource consumption during other states

Inventive Principle:

Principle #15Dynamics

3Adaptability or versatility

If data is transmitted to voice assistant service for keyword recognition, then comprehensive command recognition is achieved, but processing time increases and privacy is reduced

Engineering Contradiction:

Improvekeyword recognition capabilityVSAvoidcommand processing time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent extracts the command keyword recognition function from the cloud-based voice assistant service and implements it locally in the command keyword engine. This extraction allows the system to recognize playback commands locally without transmitting audio data to external services, eliminating transmission delays and reducing processing time while maintaining privacy

Inventive Principle:

Principle #2Taking out (Extraction)

4Measurement precision

If phonetically similar words are detected, then wake-word sensitivity is improved, but false wake word triggers increase

Engineering Contradiction:

Improvewake-word detection sensitivityVSAvoidfalse wake word triggers

Core Design Contradiction:

Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The system applies local quality by giving different detection characteristics to different keyword engines. The wake-word engine maintains high sensitivity for service invocation, while the command keyword engine uses localized playback condition checking to filter out false triggers, allowing each component to have optimized detection quality for its specific purpose

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS11854547B2Network microphone device with command keyword eventing

Publication Date: 2023.12.26 SONOS INC

AI summary

In one aspect, a playback device includes a voice assistant service (VAS) wake-word engine and a command keyword engine. The playback device detects, via the command keyword engine, a first command keyword of in voice input of sound detected by one or more microphones of the playback device. The playback device determines an intent based on at least one keyword in the voice input via a local natural language unit (NLU). After detecting the first command keyword event and determining the intent, the playback device performs a first playback command corresponding to the first command keyword and according to the determined intent. When the playback device detects, via the wake-word engine, a wake-word in voice input, the playback device streams sound data corresponding to at least a portion of the voice input to one or more remote servers associated with the VAS.