Voice Authentication Using Physical Variables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing devices face challenges in securely differentiating between authorized and unauthorized verbal user inputs in public or social environments, where unintended speech from others can trigger actions, and malicious recording of commands can lead to unauthorized operations.
Innovation Solution
The use of sensors to detect physical variables associated with the authorized user, such as pulse rate, mouth movement, and nerve impulses, combined with machine-learning techniques to generate confidence values and authenticate user inputs, ensures that only intended speech is recognized and acted upon.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If speech recognition is enabled in public environments, then device responsiveness to user commands is improved, but unauthorized operations from nearby speech increase
Solution Approach 1:
The system segments the speech authentication process into multiple independent verification stages: initial speech detection, confidence value generation, threshold comparison, and optional secondary authentication. This segmentation allows the device to respond quickly to legitimate speech while systematically filtering out unauthorized commands through progressive verification layers.
Solution Approach 2:
The patent introduces confidence values as an intermediary mechanism between speech detection and command execution. Instead of directly executing commands from detected speech, the system first generates and evaluates confidence values to determine whether the speech originates from an authorized user, thereby mediating between responsiveness and security requirements.
2Reliability
If multiple authentication methods are used to verify speech authenticity, then security against unauthorized operations is improved, but system complexity increases
Solution Approach 1:
The system implements partial authentication by using confidence value thresholds that require only a portion of possible verification criteria to be met. Instead of demanding all possible authentication factors, the system accepts speech commands when confidence values exceed predetermined thresholds, providing sufficient security without requiring excessive verification steps for every command.
Solution Approach 2:
The authentication system dynamically adjusts its verification requirements based on contextual factors such as the type of command being issued, the user's historical authentication patterns, and environmental conditions. More critical commands trigger stricter verification, while routine commands use simpler confidence-based authentication, allowing the system to optimize between security and complexity in real-time.
Data Source
AI summary
A computing device senses speech or other user input. One or more physical variables pertaining to a user of the computing device are also sensed, and respective signals are analyzed or compared to the user input. The analysis determines if the user input is likely that of an authorized user, and assigns a confidence metric to that determination. The computing device may then perform actions corresponding to the speech or user input content in accordance with the determination.