Research On Adaptive Concurrent Transmission Control Technology For Heterogeneous Wireless Links

With the continuous emergence of various high-bandwidth services such as high-definition video live broadcast,telemedicine,and unmanned driving,people have also put forward higher requirements for the transmission rate of the network.Nowadays,it is difficult for a single network to meet the business needs of different users anytime,anywhere.Therefore,using heterogeneous wireless networks to realize multi-path concurrent transmission has become one of the important research methods to solve this problem.However,in a heterogeneous wireless network environment,the conditions of different links are complex and changeable,coupled with the mobility of terminals,multi-path concurrent transmission will inevitably lead to head-of-line blocking and out-of-order receivers,which seriously reduces transmission efficiency.Therefore,aiming at the problems existing in multi-path transmission,this thesis deeply studies the adaptive concurrent transmission control technology of heterogeneous wireless links.Based on the introduction of the existing multi-path concurrent transmission control algorithms,this thesis mainly studies the following three aspects:(1)A multi-path concurrent transmission control algorithm based on adaptive network coding is proposed.On the basis of analyzing the problems of multi-path transmission,the algorithm introduces A3 C reinforcement learning,through adaptive network coding,according to the current network conditions,intelligently selects the encoding packet size and redundancy size,so as to solve the problem of out-of-order data packets.The simulation results show that,compared with the multi-path algorithm using pipeline network coding,the algorithm can improve the transmission rate by about 10%,thereby effectively improving the user experience.(2)A adaptive congestion control algorithm based on model and data(Model and Data Adaptive Congestion Control,MADACC)is proposed.After analyzing the advantages and disadvantages of the traditional congestion control algorithm and the learning-based congestion control algorithm,the algorithm fully combines the advantages of them,and divides each decision cycle into four cycles.They are exploration period,transition period,evaluation period and decision-making period.In the decision-making period,the algorithm selects the optimal sending rate based on the utility evaluation function value according to the preferences of different services.The simulation results show that the algorithm solves the problems of poor adaptability of traditional congestion control algorithms and high overhead of learning-based congestion control algorithms,and effectively improves the adaptability and practicability of the algorithm.(3)Design and implement a multi-stream transmission prototype verification system for large-bandwidth services.The system provides a larger access bandwidth for node access through multiple networks such as 4G,5G and Wi Fi coordinated transmission modules.At the same time,four adaptive scheduling strategies are designed according to the large-bandwidth business requirements,namely the Overflow strategy,the Priority strategy,the Least strategy and the Weight strategy.The system can flexibly select a matching transmission strategy according to different business requirements,thereby effectively improving the user experience when transmitting large-bandwidth services.