智能语音机器人前端语音处理系统的设计及实现

Design and Implementation of Front End Speech Processing System for Intelligent Voice Robot

语音前端处理是智能语音机器人中一项关键的技术。传统的语音前端处理方法大多通过DSP来实现,大大增加系统复杂度和研发成本。基于WebRTC语音库,采用SRP-PHAT声音定位算法实现语音前端处理系统可以解决上述问题。该系统使用C++语言实现,可以直接部署在普通通用处理器或嵌入式ARM处理器中。经过实际功能及性能测试,该系统可以满足智能语音机器人对前端语音处理的要求。

Speech front-end processing is a key technology in intelligent speech robot.The traditional DSP implementation method will increase the system complexity and cost.This paper presents a front-end speech processing system based on WebRTC speech database and SRPPHAT sound localization algorithm.The system is implemented by C++voice and can be directly deployed in general purpose processor or embedded ARM processor.After the function and performance test,the system can meet the requirements of intelligent voice robot for front-end voice processing.