Active Queue Management Of Wireless Access Network

In recent years, with the rapid development of wireless technology, the access network has been applied to people's daily lives. However, congestion, which affects the global performance, is still the key issues. How to avoid network congestion, is a currently research hotspot of wireless access network. Congestion control's significance lies not only in making a fair distribution of network resources, but also making the network coped with more complex environment.However, because of its own characteristics of wireless access network, traditional congestion control does not work well. To guarantee QoS, it is important to design new congestion control algorithms for wireless access network.This dissertation focuses on the wireless access network congestion, conducts further research on the characteristic and shortcomings of active queue management, and proposes a number of new congestion control algorithms. Simulation results the effectiveness of these proposed algorithms. The main concepts and contributions are lies below:(1). To solve REM congestion detection shortcoming caused by price production mechanism, a new price-based congestion control strategy-PREM is proposed. Different from REM, this algorithm regards the flow rate change as the congestion signal, discards queue control as a primary goal, and introduces the queue buffer to the price that makes it adapt to the change of wireless environment. Simulation results show that, PREM is effective in reducing packet loss and promoting transmission efficiency.(2). Base on the slow response of REM algorithm in the wireless access network, a new congestion control method based on sensitive queue-SQREM is proposed. In order to get faster response, the proposed algorithm uses square of queue error as a control input to replace queue error. and use single neural network to optimize the parameters. Simulation results show that, SQREM is easy to implement and good in self-learning.(3). On account of the two problems of wireless access network:①The wireless access network does not have an accurate TCP/AQM traffic models;②Traditional congestion control algorithms has poor robustness in a wireless environment,a double-layer congestion control method based on fuzzy and REM-FRC is proposed. The algorithm uses switching control method. When a sudden change of data filow happens in router, fuzzy controller will work and enhance the response speed; When data flow is gently, REM controller will work and maintain the queue stability. Simulation results show that, FRC is a robustness AQM scheme with fast response. Finally a comprehensive overview of the dissertation was summarized and points out the future research directions.