Robust congestion control for TCP/AQM using integral backstepping control

Predictive sliding‐mode congestion control for wireless access networks with singular and non‐singular control gain

“…Analytical results demonstrated that the procedure can effectively stabilise the system and can control the queue length to the desired value through the control input signal (26). To illustrate the efficacy, the proposed procedure is compared to some other procedures such as IBS [27], which is a non‐linear robust procedure for the networks with the non‐singularity condition, finite‐frequency AQM (FFAQM) [13], which is proposed robustly to a new AQM model with time‐varying capacity, proportional–integral (PI) [34] and random early detection (RED) [34], which are famous protocols in communication networks. Since the network model, which is considered in [13] to design the procedure, is different from the assumed model in this study, FFAQM is compared to the others via NS2, which operates as a real network.…”

“…Remark 1 The RTT value and number of the users can be considered to be constant in a relative long time, refer to [6, 8, 9, 27, 30, 33, 34]. Moreover, the bottleneck link capacity is assumed to be a nominal value () together with variations caused by fading effects in a wireless access network ().…”

“…The input delay is ignored in most of the papers such as [25, 26]. Integral backstepping (IBS) [27, 28], asynchronous model predictive control [29], and adaptive fuzzy funnel control [30] are proposed for a communication network with a non‐linear model. However, the input delay is also neglected in them.…”