Vehicular Speech Biasing via Sensor Data

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Automated assistants in vehicular environments often misinterpret spoken utterances related to vehicle operations due to lack of environmental context, leading to incorrect actions and increased user input, which wastes computational resources.

Innovation Solution

Implementing a biasing system that uses vehicle sensor data to determine whether to prioritize a user manual corpus of data specific to the vehicle or a broader corpus of data when processing spoken queries, based on factors like temporal and linguistic relationships, user familiarity, and explicit indications, to provide more accurate and context-specific responses.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the automated assistant searches over both the first corpus of data (vehicle-specific user manual) and the second corpus of data (broader web-based corpus), then the comprehensiveness of information retrieval is improved, but the computational resource consumption increases

Engineering Contradiction:

Improveaccuracy of interpretationVSAvoidcomputational resource consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent segments the search process into two phases: first searching the vehicle-specific corpus (user manual) for contextually relevant information, then conditionally searching the broader corpus only when necessary. This segmentation reduces unnecessary computational resources while maintaining comprehensive information retrieval when needed.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies partial action by performing only the necessary search scope - searching the vehicle-specific corpus for all queries, and only searching the broader web-based corpus when the query cannot be sufficiently answered by vehicle-specific information or when explicitly required. This avoids excessive computational action while maintaining reliability.

Inventive Principle:

Principle #16Partial or excessive action

2Adaptability or versatility

If the automated assistant searches over the broader corpus of data without biasing towards vehicle-specific data, then the versatility of information sources is improved, but the relevance of responses to vehicular context deteriorates

Engineering Contradiction:

Improveinformation source diversityVSAvoidrelevance to vehicular context

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies local quality by making the search strategy context-dependent: for queries related to vehicle operations, it prioritizes the vehicle-specific user manual corpus; for general queries, it may incorporate the broader web-based corpus. This ensures that each search operation has the appropriate quality and focus for its specific context.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent introduces a biasing mechanism as an intermediary that mediates between the two corpora. This biasing system analyzes the query context and determines the appropriate weight to give to vehicle-specific versus general information sources, ensuring that vehicle-specific relevance is maintained while still allowing access to broader information when appropriate.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Device complexity

If the automated assistant does not use vehicle sensor data to bias the search, then the simplicity of the processing system is maintained, but the precision of query interpretation deteriorates

Engineering Contradiction:

Improveprocessing system complexityVSAvoidquery interpretation accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies preliminary action by obtaining vehicle sensor data and analyzing query context before performing the search. This preliminary analysis allows the system to bias the search towards the appropriate corpus based on current vehicle state, improving interpretation accuracy without requiring complex real-time adjustments during the search process itself.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS20250006202A1Biasing interpretations of spoken utterance(s) that are received in a vehicular environment

Publication Date: 2025.01.02 GOOGLE LLC

AI summary

Implementations described herein relate to various techniques for biasing interpretations of spoken utterances that are received in a vehicular environment. For example, implementations can receive a spoken utterance that includes a query from a user of a vehicle and obtain a corresponding vehicle sensor data instance generated by vehicle sensor(s) of the vehicle. Some implementations can determine to execute a search over only a first corpus of data, but not a second corpus of data, to obtain a given response to the query based on various criteria, including at least the query, the corresponding vehicle sensor data instance, a corresponding timestamp associated with the corresponding vehicle sensor data instance, and/or a corresponding duration of time the user has been associated with the vehicle. Additional, or alternative, implementations can execute a search over both the first and second corpora of data, and obtain the given response based on the criteria.