Design of state feedback controller based on state-dependent delay modeling for congestion control in internet

Abstract

Based on the fact that the queuing delay involved in the congestion control algorithm is state-dependent, the model of communication over TCP network protocol is transformed to a state-dependent delay differential equation. Hence, our paper firstly presented a sufficient condition for the stability of this class of systems. The results are presented in terms of linear matrix inequality (LMI). After, the general results obtained are applied to design an Active Queue Management (AQM) scheme for congestion control in internet. The proposed method is verified by simulation examples using NS software and the effectiveness and superiority of our method with respect to other control schemes such as PI and RED is shown.

Guaranteed Cost Control of Networked Control Systems Under Transmission Control Protocol with Active Queue Management

“…τ max [25] 0.0283 [26] 0.2105 [27] 0.1062 [28] 0.4927 [29] 0.1728 [30] 0.3105 [31] 0.0902 Method proposed in this paper 0.6333 of the guaranteed cost function) can be obtained as J Ã 0 ¼ 28:0783 . The state response of NCS is shown in Fig.…”

“…As shown, the networked control system becomes steady at 30s. Now, we use the method proposed in [25][26][27][28][29][30][31] to investigate the same system under consideration of the effect of the delay. The maximum allowable upper bounds of the transmission delay that guarantee the stability of a system controlled over a network are given in Table I.…”

“…Moreover, a co‐design of both the plant controller and some of the communication network parameters has been considered . In a recent work , the dependency of the delay to the queue length was considered for designing AQM for the TCP communication in order to improve the performance of the network. With these approaches, the interaction between the control and communication systems has been taken into account.…”

A dynamic anti-windup AQM for congestion control in internet

“…Furthermore, as in [4], one can also notice that the rate of mismatch is the rate at which the queue length grows when the buffer is nonempty. Therefore, we can approximate it by δ q T where 1 T is the sampling frequency.…”

Discrete-time robust H∞ congestion control in Internet

“…A time-domain H design method was presented in [15], but the disturbance was not considered. An H congestion controller with disturbance was presented in [16], however, the disturbance considered was assumed to be linear.…”