Research On Congestion Control Algorithm Of Wireless Network Multimedia

With the rapid development of wireless communication technology, the applicationsbased on the wireless network have penetrated into every aspect of people’s lives. Since theequipment can be flexibly moved and users can easily access, wireless network is favoredby the majority of users, and has become an indispensable part of the Internet all over theworld. In December2013, the MIIT (Ministry of Industry and Information Technology)officially issued the fourth generation mobile communication business licences to mobileservice suppliers, which marked China’s mobile communication technology was enteringthe4G era. And a variety of WLAN technology increasingly matured. Therefore wirelessnetwork bandwidth was unprecedentedly upgraded, and the wide application of multimediaservices based on wireless network was heavily promoted, especially the real-timeaudio/video transmission business in wireless network. The development of wirelessmultimedia service puts forward a new challenge to the traditional transmission controlprotocol based on wired network. The wired link has stable performance and enoughbandwidth. However, in the wireless network environment, the wireless channel issusceptible to the environment and noise interference, the characteristics of the channel ismore complex. Therefore, we cannot simply apply the traditional transmission controlprotocol to the wireless network, when we study wireless network multimedia transmission.In this paper, the final purpose of the research is how to overcome the complexcharacteristics of wireless network, perfect the performance of the wireless networktransmission control algorithm, improve the quality of service of wireless multimediaservices. The main research work of this paper includes the following aspects:The paper studied the wireless network and the technical features of multimediaservices based on wireless network, analyzed several commonly used strategy of wirelessnetwork congestion control. Because of the algorithm is simple and consistent with thetraditional protocol semantics, the congestion control strategy based on the end-to-end hasbeen studied widely. We did many in-depth researches on several classic end-to-endcongestion control algorithms based on explicit loss differentiation, which providedreferences for the new algorithm in this paper. The article made an intensive study of traditional TFRC algorithm, analyzed twodeficiencies of the algorithm: First, at the end of the slow start, the slow response causesburst loss; second, in the phase of congestion avoidance the sending rate is susceptible toRtt bias. According to these two aspects, the paper put forward the improvement strategyrespectively, and verified the performance improvement of Improved TFRC by simulation.Through analysis on the advantages and disadvantages of the existing lossdifferentiation algorithms, this article proposed a new loss differentiationstrategy——ZigZag with Border Region Correction(Z-BRC), which absorbed the merits ofdifferent strategies, and made up for the deficiency of the traditional algorithm. Thealgorithm distinguished packet loss type preliminary, using one-way transmission delayROTT and packet loss number n, and divided the loss easy to be misjudged in the pendingarea, then according to ROTT, n and IAD judged the loss type of this part comprehensively.By this kind of diversified parameters algorithm, the network could judge the loss typemore accurately, and understand the degree of congestion more precisely. The algorithm issuitable for complex network structure and variable wireless network environment. Thepaper integrated Z-BRC loss differentiation algorithm into the improved TFRC, proposed anew congestion control strategy used in wireless multimedia network.In the simulation environment of NS2, this article conducted many experiments tocompare and validate the performance of the improved algorithm. The experiments wereconducted in different channel bit error rate and signal flow respectively to close to thepractical network as possible.Experiments show that, the proposed algorithm has different degrees of improvementcomparing with classical algorithms in network congestion loss rate, loss differentiationaccuracy and wireless bandwidth utilization. To some extent, makes up for the traditionalalgorithms’shortcomings in these respects.