The Research Of Mobile Network Voice Quality Problem Detection And Prevention Mechanism

With the rising mobile communications network size and networkcomplexity, the continuous improvement of the network adjustment, thescale of the growing network voice quality complaints has increasedsignificantly over the same period last year, increased by approximately20%in August2009to October, while the corresponding traditionalmonitoring circuit one by one can no longer meet the requirements ofrapidly locate the fault, there is an urgent need to find a quick and easymeans of fast and accurate positioning of abnormal call circuit, to meetincreasing user requirements, to provide users with high-qualitycommunications network.In mobile communication network, voice quality problems are mostcommonly encountered in four categories: one-way mute call, echo,crosstalk, no voice, etc. In the network daily maintenance, regarding thesearch or avoid one-way mute call, and crosstalk phenomenon, usually bythe way one by one monitoring circuit, which often possible to find outthe one-way mute call or crosstalk questions to monitor thousands ofcircuit, spent several days of time, seriously affecting the fault locationtime, resulting in a potential voice quality issues to decline of costomers’satisfaction. At the same time, crosstalk happening relatively complex dueto different one-way mute call phenomenon, string of things happenrelatively complex, fault phenomena many and varied, and there are a lotof emergent situation, probably it is judged as a normal normalconversation but in fact it is in fault, and failed to be monitored to, sowith the traditional search voice quality monitor circuit headed thequestion way and put forward a new challenge. The traditional one by onelisten circuit methods can not meet the requirements of the companies and users on the network, and hope that the scientific research results canbreak through to the limitation of the traditional monitoring methods,focusing on the development of the industry of the future (due to networkIP-based,2M circuit will gradually withdraw from stage), to explore thegeneral rules of voice quality problem, so as to find the appropriate meansof settlement, and use scientific means to prove the correctness andapplicability.In this paper, the user calling behavior is as the starting point,proposed the network voice quality "triangle knowledge model", used thedata mining method through one-way mute call, crosstalk, echo, novoice,the phenomenon of user behavior characteristics, the formation ofthe rule set, and verify its’ effectiveness; taked the rules in core networkcommunication detail records, signaling, statistics, alarm, bureau data andother information carriers, to take the technical feature mapping and getthe rules in the carrier technology expression; put the different scenes andthe the information carrier on the rules to diagnose a specific user voicequality problem; and by means of programming tools quickly locate thevoice quality problems sticking point.