In-Vehicle Noise Pattern Learning for Voice Applications

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing in-vehicle noise suppression techniques are limited in scope as they primarily focus on intended noise sources, such as a user's voice, and fail to dynamically account for changes in noise sources over time, leading to ineffective noise suppression in diverse vehicle and environmental conditions.

Innovation Solution

The implementation of in-vehicle noise-pattern learning using static and dynamic input signals from multiple sources to create a noise pattern learning model. This model embeds multimodal data, including vehicle and environmental information, to accurately and dynamically identify noise types using an adaptive time window and inter-quartile range calculations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing noise suppression techniques focus on intended noise sources like user voice, then the user voice inflections are improved, but other environmental noises and vehicle-specific noises are not effectively suppressed

Engineering Contradiction:

Improvenoise suppression accuracy for user voiceVSAvoidcoverage of noise types suppressed

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent segments noise into multiple categories including user voice, environmental noises, and vehicle-specific noises. Each noise type is processed separately through dedicated filtering paths, allowing precise suppression of each category without interfering with the others. This segmentation enables the system to maintain high accuracy for user voice while simultaneously addressing diverse noise sources.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The noise suppression system is designed with multi-functionality to handle multiple noise types simultaneously. It processes user voice inflections, environmental noises, and vehicle-specific noises through a unified framework that adapts to different noise conditions, making the system versatile across various driving scenarios while maintaining precision for the primary target of user voice.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Device complexity

If static noise suppression schemes are used, then the system is simpler to implement, but it fails to account for changes in noise sources over time

Engineering Contradiction:

Improvenoise suppression system complexityVSAvoiddynamic noise adaptation capability

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic noise suppression by continuously analyzing noise characteristics and adapting filtering parameters in real-time. The system monitors changes in noise sources over time and adjusts its suppression strategy accordingly, transitioning from static to dynamic operation to maintain effectiveness as noise conditions evolve during vehicle operation.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system incorporates feedback mechanisms that continuously monitor noise patterns and user voice characteristics, then adjust suppression parameters based on this feedback. This closed-loop approach allows the system to learn from changing conditions and adapt its behavior dynamically while maintaining a relatively simple overall architecture through efficient feedback processing.

Inventive Principle:

Principle #23Feedback

3Ease of manufacture

If one-size-fits-all noise suppression is applied, then the system is easier to implement, but it lacks capability to account for specific vehicle and road type noises

Engineering Contradiction:

Improvesystem implementation easeVSAvoidvehicle-specific noise identification accuracy

Core Design Contradiction:

Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent applies local quality by tailoring noise suppression parameters to specific vehicle types and road conditions. Instead of uniform processing, the system adjusts filtering characteristics locally based on the detected noise profile, vehicle specifications, and road type, enabling precise suppression of vehicle-specific noises while maintaining ease of implementation through automated parameter selection.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250029598A1Fine-grained in-vehicle dynamic noise pattern learning for voice applications

Publication Date: 2025.01.23 GM GLOBAL TECHNOLOGY OPERATIONS LLC

AI summary

Aspects of fine-grained, dynamic noise pattern learning for voice applications include a vehicle, the vehicle having a body with a cabin. Embedded within the vehicle is a processor coupled to memory. The processor may be configured to embed multimodal data for environment and vehicle data. The embedded acoustic data may be from microphone-captured data in the cabin. The processor may concatenate the embeddings to form a latent vector characterizing the embeddings to thereby estimate a mean and variance of the latent vector using an adaptive time window. The processor may identify a noise type using the mean and variance of the latent vector, the noise type identification being fine-grained via the adaptive time window to accurately emulate vehicle noise.