一种便携式无线语音会议系统设计及实现

发布时间：2018-02-13 12:37

本文关键词： Ad-hoc 语音会议系统嵌入式 VoIP　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：对讲机是一种便携式无线半双工语音通信设备,被广泛应用于生产生活中。但是对讲机是一种半双工通信系统,使用对讲机进行发言时需要按下发送按钮。但是,在进行野外工作时,会发生工作人员双手正在完成工作而无法按下发送按钮机进行发言的情况。基于有线网络的语音会议系统是一种能够进行多方全双工语音通信的系统,但是存在设备便携性差和需要有线网络连接的缺点。为满足野外工作环境下进行语音通信的需求,需要一种便携式无线语音会议系统。本文提出了一种基于ad-hoc网络的便携式无线语音会议系统设计方案。系统由多个对等节点通过ad-hoc网络连接而成。节点基于嵌入式系统实现以减小系统体积,硬件平台选用ARM处理器S5PV210,软件平台选用嵌入式Linux。软件系统按功能的不同被分为3个模块：网络协议模块、语音传输模块、按键输入模块。网络协议模块实现了自行设计的ABP(Ad-hoc Bootstrap Protocol)协议,解决了ad-hoc网络中的IP地址分配问题,还实现了节点上线通知、在线通知和下线探测功能；语音传输模块实现了语音数据的采集、发送、接收和回放。语音信号在经过麦克风采集后进行量化,然后进入编码器进行编码,编码后的数据通过UDP组播在IP网络中传输,收到数据节点将数据解码后与来自其它节点的数据混音,然后通过耳机回放；键输入模块读取按键输入并传给其它模块。本文采用多进程模型、使用函数库对各个模块进行了实现,并采用UNIX domain socket作为进程间通信设施实现进程间通信。网络协议模块基于事件循环的高性能网络编程库libuv实现,网络协议中的所有通信均使用UDP广播进行。语音传输模块使用功能强大的流媒体库mediastreamer库实现。语音编码选择了编码质量较高的speex编码器。为了取得更好效果,使用了speex编码器的回声消除、语音活动性检测、丢包补偿功能。本文在E8开发板上对系统进行了初步测试。测试结果表明系统能够实现多对多全双工无线语音通信,通信距离可达120m,可进行3点语音会议,语音总体清晰流畅。
[Abstract]:Walkie-talkie is a portable wireless semi-duplex voice communication device, which is widely used in production and life. But the interphone is a half-duplex communication system. In the field work, there will be a situation where the staff are doing the work with both hands and can't press the pushbutton machine to speak. The voice conference system based on the wired network is a system that can carry out multiparty full-duplex voice communication. However, there are some disadvantages such as poor portability of equipment and the need for wired network connection. In order to meet the needs of voice communication in field work environment, A portable wireless voice conference system is needed. This paper presents a design scheme of portable wireless voice conference system based on ad-hoc network. The system is composed of multiple peer nodes connected by ad-hoc network. The nodes are embedded. To reduce the volume of the system, The hardware platform uses ARM processor S5PV210and the software platform adopts embedded Linux. The software system is divided into three modules according to its function: network protocol module, voice transmission module, voice transmission module. The key input module. The network protocol module realizes the self-designed ABP(Ad-hoc Bootstrap protocol, solves the IP address assignment problem in the ad-hoc network, also realizes the node on-line notification, the on-line notification and the offline detection function. The speech transmission module realizes the acquisition, transmission, reception and playback of voice data. The voice signal is quantized after the microphone is collected, and then entered into the encoder for coding. The encoded data is transmitted through UDP multicast in IP network. The received data node decodes the data and remixes the data with the data from other nodes, then replays the data through headphones; the key-input module reads the key-press input and passes it to other modules. The function library is used to implement each module, and UNIX domain socket is used as the inter-process communication facility to realize inter-process communication. The network protocol module is implemented by libuv, a high-performance network programming library based on event cycle. All the communication in the network protocol is carried out by UDP broadcast. The speech transmission module is realized by using the powerful streaming media library mediastreamer library. The speech coding selects the speex encoder with high coding quality. The echo cancellation, voice activity detection and packet loss compensation function of speex encoder are used. The system is tested on E8 development board. The test results show that the system can realize many-to-many full-duplex wireless voice communication. The communication distance can reach 120 m, can carry on 3-point voice conference, the speech is clear and smooth overall.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN912.3

【相似文献】