Multi-Assistant Acoustic Event Detection With Cross-Device Validation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing techniques for detecting acoustic events in an ecosystem of assistant devices are prone to false positives, leading to unnecessary activation of automated assistant components and resource wastage.

Innovation Solution

Simultaneously process audio data from multiple assistant devices using event detection models to verify the occurrence of acoustic events, reducing false positives by ensuring temporally corresponding audio data is processed across devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If acoustic event detection is performed at a single assistant device, then the system responds quickly to detected events, but false positive detection rate increases

Engineering Contradiction:
Improveacoustic event detection accuracyVSAvoiddetection system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent combines acoustic event detection results from multiple assistant devices by processing temporally corresponding audio data captured simultaneously at different devices. This merging approach cross-validates detections across devices, reducing false positives while maintaining individual device simplicity.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If multiple assistant devices process audio data simultaneously, then false positive detection is reduced, but computational resources are consumed

Engineering Contradiction:
Improveacoustic event detection accuracyVSAvoidcomputational resource consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary filtering by identifying temporally corresponding audio data segments across devices before full processing. Only audio data captured during the same time window is processed together, reducing the total computational load while maintaining detection reliability.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts and processes only the relevant temporal portion of audio data from multiple devices - specifically the time windows where acoustic events are detected. This extraction approach avoids processing entire audio streams, reducing computational resource consumption.

Inventive Principle:
Principle #2Taking out (Extraction)

3Reliability

If acoustic event detection is performed across multiple devices, then false positives are reduced, but network communication overhead increases

Engineering Contradiction:
Improveacoustic event detection accuracyVSAvoidnetwork communication energy
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The system extracts and transmits only the essential detection information - temporal markers and audio data segments - from multiple devices rather than exchanging complete audio streams. This reduces network communication overhead while enabling cross-device validation.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP4042414B1Simultaneous acoustic event detection across multiple assistant devices
Publication Date: 2025.11.19 GOOGLE LLC
  • EP4042414B1 patent drawingFigure 1
  • EP4042414B1 patent drawingFigure 2A
  • EP4042414B1 patent drawingFigure 2B

AI summary

Implementations can detect respective audio data that captures an acoustic event at multiple assistant devices in an ecosystem that includes a plurality of assistant devices, process the respective audio data locally at each of the multiple assistant devices to generate respective measures that are associated with the acoustic event using respective event detection models, process the respective measures to determine whether the detected acoustic event is an actual acoustic event, and cause an action associated with the actional acoustic event to be performed in response to determining that the detected acoustic event is the actual acoustic event. In some implementations, the multiple assistant devices that detected the respective audio data are anticipated to detect the respective audio data that captures the actual acoustic event based on a plurality of historical acoustic events being detected at each of the multiple assistant devices.