用于无线智能音频设备的片上系统和无线处理方法
By using a system-on-a-chip with a large-core-small-core collaborative configuration, the low-power first processor core is used for initial voice screening and RNN network recognition, while the high-power second processor core is activated for keyword recognition. This solves the problem of excessive power consumption in wireless smart audio devices and achieves low-power and high-efficiency keyword recognition.
CN116705033BActive Publication Date: 2026-07-17BESTECHNIC SHANGHAI CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BESTECHNIC SHANGHAI CO LTD
- Filing Date
- 2023-05-17
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing wireless smart audio devices' on-chip systems consume excessive power when performing voice command detection, making it difficult to achieve efficient and low-power keyword recognition.
Method used
The system-on-chip adopts a big-core-little-core collaborative configuration. It uses the low-power first processor core for initial screening of human voices, recognizes human voices through a trained RNN network, and wakes up the high-power second processor core after recognizing human voices to perform echo cancellation and automatic speech recognition.
Benefits of technology
It achieves low-power, high-efficiency keyword recognition, avoiding excessive overall power consumption of the on-chip system and reducing system cost.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116705033B_ABST
Abstract
本申请涉及一种用于无线智能音频设备的片上系统和无线处理方法。所述片上系统包括第一处理器核和第二处理器核，所述第一处理器核的主频频率低于第二处理器核且功耗低于第二处理器核。第一处理器核获取音频数据；以及利用训练好的RNN网络来识别人声；在识别出人声的情况下，利用所述核间通信模块开启到核间通信，以便传输音频数据并使得所述第二处理器核唤醒。第二处理器核仅在唤醒情况下接收音频数据，并对所接收的音频数据进行AEC，并对除噪后的音频数据进行ASR，从而识别用户语音命令的关键词。如此，能够对不时出现的用户的语音命令中的关键词进行精确且及时的检出，避免片上系统的功耗过大，并且还能有效控制片上系统的成本。
Need to check novelty before this filing date? Find Prior Art