中文摘要：

针对TCP（Transmission Control Protocol,传输控制协议）网络的拥塞控制问题,基于T-S（Takagi-Sugeno）模糊模型,采用滑模控制理论提出了一种新的AQM（Active Queue Management,主动队列管理）算法。考虑到TCP网络中存在的不确定和时变时滞因素,首先利用T-S模糊模型对网络进行建模,然后利用线性矩阵不等式设计了一个渐近稳定的滑模面,而且还给出了一种能够明显减小滑模面附近抖振的趋近律,基于该趋近律设计的控制律能够有效地抑制路由器中队列长度的振荡,并使其快速收敛于期望值。仿真结果表明,该算法与普通的滑模控制算法相比具有更好的稳定性和鲁棒性,能够很好地适应复杂多变的TCP网络环境。

英文摘要：

As for the problem of congestion control in transmission control protocol （TCP） networks, a novel active queue management （AQM） algorithm based on Takagi-Sugeno （T-S） fuzzy model was put forward by employing sliding mode control theory. Considering uncertainties and time-varying delays in networks, a network model was obtained by using T-S modeling method. An asymptotically stable sliding surface was designed by linear matrix inequality （LM1）. In addition, a novel reaching law that could obviously reduce chatter around the sliding surface was put forward. Based on the reaching law, the control law was designed to constrain oscillation of the queue length in router effectively. And the queue length converged to the desired value quickly. Simulation results demonstrate that the stability and robustness of the algorithm is better than that of common sliding mode control algorithm, so it is very suitable for the mutable TCP network environment.