Research On Concurrent Multipath Transmission For Real-time Videos Over Heterogeneous Wireless Networks

With the rapid development of mobile intelligent terminals and wireless network technologies,mobile users can enjoy various services brought by mobile internet anytime and anywhere.Currently,video program has become the main service for mobile users.Limited by the current mobile network bandwidth,non-real-time video are the main service at present.Real-time videos have high requirements of bandwidth over wireless networks.It is obviously that single wireless network with limited bandwidth cannot provide users with low-latency real-time video services.As the increasement of users,wireless networks will not be able to meet the growing demands of users,even though in the 5G networks.Fortunately,heterogeneous wireless networks will be the general trend of future networks.Existing research has confirmed that by converging different network resources and providing users with services concurrently,transmission bandwidth can be significantly improved.Therefore,it is of great significance to research real-time video transmission technology over heterogeneous wireless networks.In recent years,more and more scholars have focused on video transmission technology over heterogeneous wireless networks,especially on concurrent multipath transmission.Existing researches shown that the efficiency of concurrent multipath transmission over heterogeneous wireless networks can be improved by a reasonable distribution strategy.However,mobile terminals will pay a higher cost,such as power cost or tariff cost,while connecting to multiple network simultaneously.Therefore,how to get a tradeoff between transmission performance and cost,and improve the user's QoE is the key to real-time video applications over heterogeneous wireless networks.In this paper,the user's QoE model on concurrent multipath transmission is established based on transmission performance and cost.On this basis,aiming at maximizing user's QoE,a video distribution strategy based on user's QoE in different scenarios is proposed.Experimental results show that the strategy can adaptively adjust the video multipath distribution scheme in different scenarios for the purpose of improving the user's QoE.On the other hand,bandwidth can be improved by concurrent multipath transmission technology.However,the time-varying characteristic of wireless networks cannot be changed,and fixed-rate videos may still suffer frame drops when bandwidth is insufficient.In this paper,a rate control method based on buffer status is proposed,which aims to minimize the end-toend delay.The proposed algorithm is based on traditional rate control method of single-path video transmission,and considering concurrent multipath transmission characteristics simultaneously.Specifically,an end-to-end delay model of video frames under concurrent multipath transmission is established,and the upper bound of video frames bits with lowlatency constraints is predicted.Then,we proposed an adaptive rate control strategy,which aims to improve video quality under the constraint of end-to-end delay.Experimental results show that the proposed algorithm can effectively guarantee low latency and improve video quality under time-varying networks.Finally,on the basis of network emulation devices,a real-time video concurrent multipath transmission testbed is built on Linux,and the above proposed algorithm is tested and verified on the testbed.Experimental results show that the proposed algorithm are not only theoretically supported,but also perform well in practical systems,which can provide a reference for practical applications.