Design And Implementation Of A Multi-party Voice Communication System Based On Ad-Hoc

Nowadays, sound communication equipment based on embedded technology have adapted to more and more application scenarios such as civilian areas, industrial areas and the military field. Because of the high cost of the traditional telecom network, new voice transmission technologies are emerging. However, the application of these new technologies are still limited because of some flaws. For example, bluetooth helmet intercom has applied to the motorcycle helmet, but it has small user capacity and low rate in communication. Equipment for web conference is expensive and requires regular maintenance, not to mention the poor mobility. Compared to Bluetooth and Zigbee, Wi-Fi technology is more suitable to realize functions of the mobile communication device. The field of using Wi-Fi for voice transmission still has much room for development. Therefore, the design of voice communication system based on Wi-Fi is of great significance.This system is based on the infrastructure-less network(Ad-Hoc). It allows full duplex and multi-party voice communication among all the members in the network. Selecting Tiny210 demoboard as the hardware development platform and choosing Linux system as the software part. At first, the systematic entirety structure and the software and hardware schemes are designed according to the project requirements. And then, the system is divided into three modules including Ad-Hoc self-network, voice communication and volume adjustment. The Ad-Hoc network module could automatically set up the multi-hop Ad-Hoc network for all members and manage them in real-time. Members' information of joining, leaving or crashing could be detected in real-time by other members, which makes the network system with strong mobility and survivability. Voice communication module, which is based on the Mediastreamer-2.9.0 library, is designed to process and control the audio stream through the operation of filters. Using the Vo IP technology combined with RTP/RTCP protocol in the part of audio stream processing to realize the functions of sending and receiving voice and real-time transmission. Using Speex encoding technique in the part of speech coding and designing the mixing synthesis functions to realize the multi-party voice communication. Designing the silence detection features to improve the quality of voice communication in the system. In the part of volume adjustment, users can push the buttons on the demoboard to transfer commands to the voice communication module to realize dynamic volume control. Using the multi-process technology to interact with each other among the various modules. Setting up the TCP SOCKET communication mode with one host and multiple clients to realize the transmission of information. In the end, the tests of the basic functions and vital indicators are shown and analyzed.The results of test show that the multi-party voice communication based on Ad-Hoc network has reached the design requirements of the project. The system can be in long and stable operation. All members can join or quit the network at any time. The members can carry out the multi-party voice communication with good voice quality. The system can through at most three brick walls or four wooden walls, and its farthest distance is up to 125 meters.