Abstract: Aiming at the dynamic nonlinear flow model of TCP in network congestion control of Internet, this paper proposes a congestion control algorithm for Active Queue Management(AQM). It designs a state observer to estimate the unmeasured state, applies the backstepping approach technique and Lyapunov direct method. Closed-loop system is asymptotically stabilized through output feedback. Simulation results show that AQM algorithm based on backstepping approach needs less tuning time, has lower packet loss rate and higher link utilization rate.

Key words: Active Queue Management(AQM), congestion control, backstepping approach, output feedback

摘要： 针对Internet网络拥塞控制中的TCP动态非线性流体模型，提出用于网络主动队列管理(AQM)的拥塞控制算法，设计用于估计未知状态的状态观测器，采用反步法技术和Lyapunov直接方法，通过输出反馈实现闭环系统的渐近稳定。仿真实验结果表明，基于反步法的AQM控制算法调整时间小、丢包率低、链路利用率高。

关键词: 主动队列管理(AQM), 拥塞控制, 反步法, 输出反馈

CLC Number: 

• TP393.07