Voice Navigation via Multi-Engine Website Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Natural User Interface (NUI) systems for navigating websites using voice commands face challenges in accurately identifying the intended website from spoken expressions, as users may pronounce URLs differently, making traditional methods cumbersome and inefficient.
Innovation Solution
A system employing multiple software engines, including a one-to-one lookup engine, a top URLs engine, and a general URL retrieval engine, to compare spoken expressions against closed set grammars and search engine query logs to determine the best matching website, using fuzzy logic and heuristics to establish confidence levels for accurate navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional user interface methods (mouse, keyboard, touch) are used to manually enter a URL, then navigation to websites can be achieved, but the operation becomes cumbersome and inefficient
Solution Approach 1:
The patent replaces mechanical input methods (keyboard typing, mouse clicking) with voice-based acoustic input. Users speak the website name or URL, and the system converts speech to text to automatically navigate to the destination, eliminating manual typing and significantly reducing the time and effort required for website navigation
Solution Approach 2:
The system automatically processes the spoken input through speech-to-text conversion, matches it against stored website data, and executes navigation without requiring user confirmation or additional manual steps. The system serves itself by completing the entire navigation task based solely on the voice command
2Ease of operation
If voice commands are used for website navigation, then ease of operation is improved, but accuracy in identifying the intended website from variable pronunciations deteriorates
Solution Approach 1:
The system pre-stores multiple variations of website identifiers including full URLs, domain names, and common shorthand versions in a database. When a voice command is received, the system compares the spoken input against this pre-prepared reference data to find the best match, enabling accurate identification despite variations in how users pronounce or phrase the website name
Solution Approach 2:
The system employs a matching algorithm that compares the speech-to-text output against stored website data and provides feedback through confidence scoring. When the match confidence exceeds a threshold, navigation proceeds automatically; otherwise, the system can request clarification or present options to the user, ensuring accurate website identification
3Measurement precision
If multiple software engines are used to match spoken expressions to websites, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the website matching task into separate functional modules: a speech-to-text conversion engine, a one-to-one lookup engine for exact matches, a top URLs engine for common website patterns, and a general URL retrieval engine for broader searches. Each engine handles specific types of queries, improving overall accuracy while maintaining manageable complexity through modular design
Data Source
AI summary
A system and method are disclosed for navigation on the World Wide Web using voice commands. The name of a website may be called out by users several different ways. A user may speak the entire URL, a portion of the URL, or a name of the website which may bear little resemblance to the URL. The present technology uses rules and heuristics embodied in various software engines to determine the best candidate website based on the received voice command, and then navigates to that website.