Vehicle Voice Intention Inference From Ambiguous Commands
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle voice recognition systems struggle with accurately interpreting ambiguous voice commands and require manual registration of commands, leading to delayed application of frequently used utterances and reduced reliability due to slight variations, and existing technologies do not effectively improve usability or recognition rates for multiple users.
Innovation Solution
A method and apparatus that utilize ambiguous voice recognition commands and subsequent action information from multiple vehicles to infer user intentions, employing machine learning to predict user intentions through pattern recognition and association rules, with feedback loops to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual registration of representative commands and variant commands is used, then the voice recognition system can provide structured command recognition, but it takes time for frequently used utterances to be applied and reduces reliability when user utterances differ slightly from registered commands
Solution Approach 1:
The system automatically learns and registers new commands by analyzing user utterances and subsequent actions, eliminating the need for manual command registration. The server autonomously updates the command database based on collected data from multiple vehicles, enabling frequent utterances to be applied immediately without manual intervention.
Solution Approach 2:
The system implements a feedback mechanism where user actions following voice commands are collected and analyzed. When a command is not recognized, the system learns from the user's subsequent manual actions to infer the intended command, then feeds this learned information back into the command database for future recognition improvements.
2Productivity
If manual review and registration of variant commands is performed, then command accuracy can be maintained, but the process is time-consuming and does not adapt to new frequently used utterances quickly
Solution Approach 1:
The system automatically performs command learning, analysis, and registration without manual intervention. The server autonomously processes collected utterance and action data to identify new commands and update the database, eliminating the need for manual command management while rapidly adapting to frequently used utterances.
Solution Approach 2:
The voice recognition server acts as an intermediary that collects data from multiple vehicles, analyzes patterns, and automatically updates command databases. This centralized intermediary process simplifies the overall system by consolidating the complex task of command learning and management in a single automated component.
3Adaptability or versatility
If conventional intention inference technology using individual user patterns is used, then specific user recognition can be improved, but it does not improve usability for other users
Solution Approach 1:
The system collects and analyzes data from multiple vehicles and users to create a universal command database that benefits all users. The server aggregates anonymized utterance and action data across the fleet, enabling patterns learned from any user to improve recognition accuracy for all users, making the system universally applicable rather than user-specific.
4Loss of information
If analysis of voice recognition failure causes is performed without correction logic, then user understanding can be improved, but there is no influence on improvement of user voice recognition rate
Solution Approach 1:
The system implements a closed-loop feedback mechanism where recognition failures are not only analyzed but actively used to improve future recognition. When a command is not recognized, the system collects the utterance and subsequent user actions, analyzes the pattern, and automatically updates the command database to prevent similar failures, continuously improving the voice recognition rate over time.
Data Source
AI summary
An embodiment computer-implemented method for predicting an intention of a user includes receiving from a vehicle first utterance data obtained by converting a voice command of the user into text, performing natural language understanding to attempt to decide the intention of the user from the first utterance data, predicting the intention of the user using stored pattern data in response to failing to decide the intention of the user, wherein the stored pattern data includes a plurality of patterns and confidence generated based on second utterance data received from each of a plurality of vehicles that are unable to decide the intention of the user and subsequent action data, generating a prompt suggesting an operation based on the predicted intention, and transmitting the prompt to the vehicle.