Dynamic ASR Feature Modification via Keyword Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Vehicle manufacturers face complexity and expense in modifying third-party provided automatic speech recognition systems, as changing or adding features requires significant effort and cannot be retroactively applied to previously manufactured vehicles.

Innovation Solution

A method that allows vehicle manufacturers to monitor speech for specific keywords to add or remove features from the ASR system at the pre-processor or post-processor software modules, enabling remote modification without retraining the entire system, thus allowing for quick changes to the functionality of ASR systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the vehicle manufacturer wants to change the functionality or features offered by the speech recognition system, then the functionality can be modified, but the process requires significant expense and complexity involving third-party re-training and re-programming

Engineering Contradiction:

Improvefunctionality modification capabilityVSAvoidmodification process complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the speech recognition system into modular components, allowing specific features to be independently added or removed through keyword-based commands. This modular approach enables the vehicle manufacturer to modify individual functionalities without requiring complete system re-training or third-party intervention, thus reducing modification complexity while maintaining adaptability.

Inventive Principle:

Principle #1Segmentation

2Adaptability or versatility

If the vehicle manufacturer wants to change the functionality of the speech recognition system, then feature changes can be made, but it cannot be applied to previously-manufactured vehicles

Engineering Contradiction:

Improvefeature change capabilityVSAvoidtime applicability to vehicle fleet

Core Design Contradiction:

Adaptability or versatilityVSDuration of action of stationary object

Solution Approach 1:

The patent implements a dynamic feature modification system where the speech recognition system can be updated in real-time through keyword-based commands. These changes are applied dynamically across the vehicle fleet without requiring physical re-training or re-programming, enabling previously-manufactured vehicles to receive functionality updates remotely and immediately, thus extending the time applicability of feature changes across the entire vehicle population.

Inventive Principle:

Principle #15Dynamics

3Adaptability or versatility

If the vehicle manufacturer wants to modify the speech recognition system, then functionality can be changed, but assistance from the third-party provider is required

Engineering Contradiction:

Improvefunctionality modificationVSAvoidmodification independence

Core Design Contradiction:

Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

The patent enables the vehicle manufacturer to independently modify the speech recognition system functionality through an automated keyword-based command system. The system self-adjusts by adding or removing features based on detected keywords, eliminating the need for third-party provider assistance. This self-service capability allows the manufacturer to autonomously adapt the system to changing requirements without external intervention.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS10083685B2Dynamically adding or removing functionality to speech recognition systems

Publication Date: 2018.09.25 GM GLOBAL TECHNOLOGY OPERATIONS LLC

AI summary

A system and method of changing features of an existing automatic speech recognition (ASR) system includes: monitoring speech received from a vehicle occupant for one or more keywords identifying a feature to remove from or add to the ASR system; detecting the keywords in the monitored speech; and adding the identified feature to or removing the identified feature from from the ASR system.