DSP-CPU Segmented Speech Wakeup for Low Power Mobile Terminals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional speech recognition systems in mobile terminals require constant CPU activity for audio processing, leading to high power consumption and reduced standby time, as they cannot enter a dormant state, and rely on external power or physical keys for wake-up, limiting mobility and battery life.
Innovation Solution
A speech recognition method and apparatus that utilizes a digital signal processor (DSP) to perform fuzzy speech recognition on audio data, waking up the CPU only when a wakeup word is detected, allowing the CPU to enter a dormant state when the word is not present, thereby reducing power consumption and extending battery life.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the CPU processes audio data continuously to enable speech wakeup function, then the speech wakeup capability is maintained, but power consumption increases and standby time decreases
Solution Approach 1:
The speech recognition system is segmented into two independent modules: a lightweight DSP module for continuous audio monitoring and a full-featured CPU module for processing only when wakeup words are detected. This segmentation allows the high-power CPU to enter dormant state while the low-power DSP maintains speech wakeup capability, resolving the contradiction between reliability and energy consumption.
Solution Approach 2:
The DSP acts as an intermediary between the audio input and the CPU. It continuously monitors audio data and only triggers the CPU when wakeup words are detected. This intermediary role allows the system to maintain speech wakeup capability without requiring the CPU to remain active, thereby reducing power consumption.
2Use of energy by moving object
If the CPU enters dormant state to save power, then power consumption decreases and standby time increases, but speech wakeup function cannot be performed
Solution Approach 1:
The system divides speech recognition functionality between DSP (continuous monitoring) and CPU (dormant state). The DSP segment handles low-power audio analysis while the CPU segment remains dormant until triggered, ensuring both power savings and maintained speech wakeup capability.
Solution Approach 2:
The DSP performs preliminary action by continuously analyzing audio data in advance before CPU activation is needed. This preliminary monitoring ensures that when the CPU needs to wake up, it can do so immediately upon detecting wakeup words, maintaining speech wakeup capability while allowing the CPU to remain dormant otherwise.
3Ease of operation
If external power supply or physical key is used for wakeup, then speech wakeup function is limited or unavailable, but mobility is affected or speech wakeup cannot be implemented
Solution Approach 1:
The system replaces mechanical wakeup methods (physical keys) with an acoustic-based DSP module that can detect speech remotely. This substitution maintains full mobility since no physical interaction is needed, while preserving and enhancing speech wakeup functionality through the always-active DSP audio monitoring capability.
Data Source
AI summary
A speech recognition method and apparatus, and a storage medium are provided. The method includes: acquiring, by a digital signal processor (DSP), audio data; performing, by the DSP, fuzzy speech recognition on the audio data; waking up a central processing unit (CPU) in a dormant state if a fuzzy speech recognition result indicates that that a wakeup word exists in the audio data. The method also includes: reading, by the CPU, data corresponding to the wakeup word in the audio data from the DSP, to obtain wakeup data; determining, by the CPU, whether the wakeup word exists in the audio data by performing speech recognition on the wakeup data; if the wakeup word exists, performing, by the CPU, semantic analysis on the audio data; and if the wakeup word does not exist, determining, by the CPU, that the fuzzy speech recognition result is incorrect and entering the dormant state.