Dual-Processor Speech Wake-Up for Handheld Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Handheld portable electronic devices face challenges in efficiently waking up from sleep mode using speech recognition due to high computational and power expenses associated with microphone occlusion detection and beamforming, especially when microphones are unpredictably occluded.
Innovation Solution
The implementation of a dual-processor system, where a primary processor handles complex tasks in wake mode and an auxiliary processor, configured for power-saving operations, detects speech commands even when microphones are occluded by processing audio signals in parallel from multiple microphones using separate speech recognition engines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If microphone occlusion detection and beamforming are used to improve speech detection accuracy, then speech recognition reliability is improved, but power consumption and computational expense increase significantly
Solution Approach 1:
The system segments speech detection functionality into two parts: a always-on auxiliary processor that performs simple wake-word detection with low power consumption, and a main processor that performs complex beamforming and occlusion detection only when needed. This segmentation allows the system to maintain speech detection reliability while significantly reducing overall power consumption during sleep mode.
Solution Approach 2:
The auxiliary processor performs partial speech detection by monitoring only for specific wake-words rather than performing full speech recognition. This partial action approach enables the system to detect wake-up commands with sufficient reliability while avoiding the excessive computational and power costs of continuous full-speech processing.
2Measurement precision
If beamforming is continuously performed to improve speech signal quality, then speech recognition accuracy is improved, but computational expense and processing time increase
Solution Approach 1:
The system segments computational tasks by having the auxiliary processor handle simple wake-word detection without beamforming, while reserving complex beamforming operations for the main processor to be activated only after a wake-word is detected. This segmentation reduces computational expense during sleep mode while maintaining speech signal quality when needed.
Solution Approach 2:
Instead of continuous beamforming, the system uses periodic action by activating the main processor and performing beamforming only after the auxiliary processor detects a wake-word. This periodic activation maintains speech recognition accuracy while significantly reducing computational expense and processing overhead.
3Speed
If the main processor remains active to detect speech commands, then wake-up detection speed is improved, but power consumption increases
Solution Approach 1:
The system segments the wake-up detection function between an always-on auxiliary processor that monitors for wake-words and a main processor that remains in sleep mode. This segmentation enables fast wake-up detection through the auxiliary processor while keeping power consumption low, as the main processor does not need to remain continuously active.
Solution Approach 2:
The auxiliary processor acts as an intermediary between the microphones and the main processor. It continuously monitors audio input for wake-words and only activates the main processor when a wake-word is detected, thereby maintaining fast response times while minimizing power consumption by keeping the power-hungry main processor in sleep mode.
Data Source
AI summary
A system and method for parallel speech recognition processing of multiple audio signals produced by multiple microphones in a handheld portable electronic device. In one embodiment, a primary processor transitions to a power-saving mode while an auxiliary processor remains active. The auxiliary processor then monitors the speech of a user of the device to detect a wake-up command by speech recognition processing the audio signals in parallel. When the auxiliary processor detects the command it then signals the primary processor to transition to active mode. The auxiliary processor may also identify to the primary processor which microphone resulted in the command being recognized with the highest confidence. Other embodiments are also described.